Poverty Global Practice Africa Region

May 2015 Mapping Poverty in Botswana 2010

Document of the World Bank

ISBN 978-99968-429-9-3

2 Mapping Poverty in Botswana 2010 Statistics Botswana Statistics Botswana Mapping Poverty in Botswana 2010 3 Abbreviations and Acronyms Contents

1 Introduction ...... 9 BCWIS Botswana Core Welfare Indicator Survey DECRG Development Research Group 2 Poverty measurement in Botswana ...... 10 EA Enumeration Area 2.1 Consumption aggregate...... 10 ELL Elbers, Lanjouw and Lanjouw 2.2 Data source ...... 10 HIES Household Income and Expenditure Survey 2.3 An Overview of Botswana’s Poverty...... 10 PDL Poverty Datum Line PSU Primary Sample Unit 3 Poverty Mapping Method...... 11 SAE Small Area Estimation 3.1 Stage zero. Comparability of Data ...... 11 SSA Sub-Saharan Africa 3.2 Stage one. Expenditure modelling...... 11 StatsBots Statistics Botswana 3.3 Stage two. Computation of the welfare measure...... 12

4 Data ...... 12

5 Results ……………………………………………………………………...... 13 5.1 Model Selection...... 13 5.2 Main Findings...... 14 5.2.1 District Poverty in Botswana...... 14 5.2.2 Village poverty in Botswana...... 14 5.2.3 Precision of poverty estimates ...... 15 5.2.4 Further checks: Application of Empirical Best Prediction...... 16

6 Conclusions...... 16

Vice President: Makhtar Diop Country Director: Asad Alam Senior Director: Ana L. Revenga Practice Manager: Pablo Fajnzylber Task Team Leader: Victor Sulla

4 Mapping Poverty in Botswana 2010 Statistics Botswana Statistics Botswana Mapping Poverty in Botswana 2010 5 List of Tables PREFACE

1 Descriptive Statistics: Dwelling characteristics (rooms, house tenure, and water source)...... 18 This is the second Poverty Map report by Statistics Botswana following the conduct of Botswana Core Welfare Indicators 2 Descriptive Statistics: Dwelling characteristics (type of toilet, lighting, cooking source, and heating)...... 19 Survey (BCWIS) in 2009/10. The first report was published in 2002/03. 3 Descriptive Statistics: Sources of income and assets...... 20 4 Descriptive Statistics: Gender, civil status, disability, and schooling of household head...... 21 The Poverty Map provides disaggregated poverty rates at lower geographical levels that the surveys are not able to 5 Descriptive Statistics: Employment and household size of household head...... 22 estimate. This is done by making use of the survey and Population and Housing Census (PFC) data. The survey data 6 Descriptive Statistics: Other socio-demographic characteristics...... 23 provides scope for in-depth poverty analysis, whilst the PHC data allows for estimation of poverty rates at lower levels 7 OLS model of the logarithm of per capita adult equivalent expenditure...... 24 because of complete coverage of the population. 8 Alpha Model...... 25 9 GLS model of the logarithm of per capita adult equivalent expenditure...... 26 The Poverty Map was compiled through the assistance of the World Bank, both financially and technically. Statistics 10 Regional Poverty: Comparison between survey and Poverty Map estimates...... 27 Botswana would like to thank the World Bank for assistance. Appreciation is also extended to Dr. Victor Sulla and Ms. Ms 11 District Poverty: Comparison between survey and census poverty estimates...... 28 Erica Rascon for the technical assistance provided in the construction of the poverty map. 12 Village Poverty: The twenty villages with lowest and highest poverty...... 29 For more information and further enquiries, contact the Directorate of Stakeholder Relations at 3671300. All Statistics List of Figures Botswana outputs/publications are available on the website at www.cso.gov.bw and at the Statistics Botswana Library (Head-Office, Gaborone). 1 Beta residuals and Fitted Values...... 42 2 Beta Model Residuals of Models 1 to 3...... 43 We sincerely thank all stakeholders involved in the formulation of this report, for their continued support, as we strive to 3 Comparison of LNEXP and per capital LNEXP models Quartile-Quartile Plot...... 44 better serve users of our products and services. 4 Comparison of per capital LNEXP models using different sets of variables Quartile-Quartile Plot ...... 44 5 Village Poverty Incidence by District Poverty Map Estimates...... 45 6 Village Poverty Density by District Poverty Map Estimates ...... 46 7 Confidence intervals of village level estimates ELL Standard Method ...... 47 8 Significant different pairwise comparisons of villages within the nation...... 47 A. N. Majelantle 9 Significant different pairwise comparisons of villages within the nation (Bonferroni Correction)...... 48 Statistician General 10 Significantly different pairwise comparisons of villages within districts (Bonferroni Correction)...... 49 May 2015 11 Proportion of villages with poverty rates significantly different from the district poverty rate (Bonferroni Correction)...... 50 12 Confidence intervals of village level estimates Empirical Bayes prediction (EB)...... 51 13 Confidence intervals of village level estimates (Only villages appearing in survey and census)...... 51

6 Mapping Poverty in Botswana 2010 Statistics Botswana Statistics Botswana Mapping Poverty in Botswana 2010 7 Acknowledgements 1. Introduction

This report was written by Ericka G. Rascon (Consultant) with the contribution of Marina Sorrentino (National Institute of The Republic of Botswana is a sparsely populated country, with 2,038,228 inhabitants living in 581,730 km2 of land of which Statistics, Italy) in data preparation and Nyaladzani Nkhwanana (GIS Cartography, Statistics Botswana) in the mapping 70 percent is dominated by the Kalahari Desert. Despite the population growth rate at a 1.9 percent, recent analysis of poverty estimates using GIS techniques. The development of the report was led by Victor Sulla (Task Team Leader and from the 2011 Population and Housing Census highlights the deceleration of Botswana’s population growth rate in recent Senior Economist, GPVDR). decades. The growth rates reported by previous censuses in 1981, 1991 and 2001 are 4.6, 3.5 and 2.4 percent respectively (Botswana Central Statistics Office, 2011). Alongside the change in population growth rates, poverty indicators have also The report benefited from extensive discussions with government officials from Botswana. The organization is grateful to changed in the last decades. The poverty rates between 2003 and 2010 decreased from 30.6 to 19.3 percent. Moffat Malepa (Labour and Poverty Statistics, Statistics Botswana) and Boago Mashadi (Statistician, Statistics Botswana). We also thank our peer reviewers, Kenneth Simler (Senior Economist, GPVDR) and Alexandru Cojocaru (Economist, Based on survey data, national and district poverty rates have been calculated using household consumption aggregates. GPVDR). Their contributions substantially improved this report. However, it is still unknown how much heterogeneity exists within the districts. To obtain poverty rates at lower geographical levels is a crucial tool for policy makers to allocate more efficiently social program transfers. The main concern of relying The poverty mapping methodology and its potential applications for Botswana were discussed with representatives from on district level poverty estimates for the identification of the poor population is that poverty rates can be explained the Government of Botswana during a workshop held in Gaborone in May 2014. This includes several units from Statistics mainly by those villages with a higher concentration of population leaving behind the least populous. Botswana and other government agencies. I extend my sincerest thanks to the following participants:, Tapologo Baakile (Director of Socio-Demographic Statistics, Statistics Botswana), S. Bakane (Department of Gender Affairs, Ministry of Labour The identification of poor areas has become one of the main keys for targeting social programs and for allocating public and Home Affairs), Kagiso Motlhabane (IT, Statistics Botswana), Bunnie Komane (Labour Statistics, Statistics Botswana), resources to finance projects. To identify the heterogeneity of poverty rates at sub-national administrative levels, several Victor Makwati (Poverty Statistics, Statistics Botswana), Malebogo Prisca Kerekang (Director of Stakeholder Relations, mechanisms have been developed. For instance, census data have been widely used for producing community profiles Statistics Botswana), Peter Mabaka (Department of Fishery, Ministry of Wildlife and Conservation), Uyapo Mosarwa (Office based on infrastructure and socio-economic characteristics to construct deprivation and unsatisfied basic need indices. of the President), Masilo Thutwa (Poverty Eradication Team), Fifi Sebolao Statistician, Statistics Botswana), Kutlwano However, the lack of detailed socio-economic information in census data, such as income or consumption, has restricted Sebolaaphuti (Office of the President), Temba Sibanda (Corporate Communication, Statistics Botswana), Gomolemo the ability to describe the variability of poverty at sub-national levels.1 Tembwe (Ministry of Local Government and Rural Development) and Tselakgopo (Ministry of Local Government and Rural Development). To overcome this problem, since the late 1990s, The World Bank Group (DECRG) and affiliated researchers have been developing methods to combine the detailed information of household surveys together with the coverage of census data. One of them is the “Poverty Map” method which aims at producing welfare indicators at a highly detailed level of spatial disaggregation.

With a growing body of applied research in country specific applications, the method is now widely recognised as a highly proficient tool for allocating social programmes at the local level. At present, around fifty countries have produced Poverty Maps with a growing number of these producing second rounds.2 As a result of this demand, an independent software designed by World Bank (DECRG) staff, has been produced to ease the computational burden of producing these maps. 3

The accuracy of the Poverty Map method is country specific. In particular, the accuracy and precision of final estimates of poverty rates rely on: a) the degree of comparability between the variables in census and survey; and b) the power of the comparable variables to predict income or expenditure in the survey. 4

This report discusses the main findings of the Poverty Map 2010 at the village level in Botswana. It also provides a discussion of how different village poverty estimates are from district poverty rates, as well as how different these are within the district. The authors also provide a detailed discussion of the precision of the Poverty Map estimates. Using the visual representation of village poverty estimates, this report highlights the villages that require urgent attention to tackle poverty. The report is organised as follows: Section 2 describes the official measurement of poverty and its evolution; Section 3 explains the World Bank method to produce small area estimates; Section 4 describes data sources and comparability of survey and census variables; Section 5 describes the model selection, main findings and precision of poverty estimates; and Section 6 concludes

1Sub-national levels refer to small areas such as localities, villages and sub-districts 2Some applications can be seen in Araujo et al. (2008); Bedi and Coudouel (2007); Hentschel et al.(2000); López-Calva et al. (2008); and Elbers et al. (2004). 3The software can be downloaded from: http://iresearch.worldbank.org/PovMap/PovMap2/ 4Henceforth, I will refer to household expenditure as the variable of interest for producing poverty estimates.

8 Mapping Poverty in Botswana 2010 Statistics Botswana Statistics Botswana Mapping Poverty in Botswana 2010 9 2. Poverty measurement in Botswana 3. Poverty Mapping Method

2.1 Consumption aggregate This section describes the Poverty Map method to produce district and village poverty estimates in Botswana. Since the late 1990s, the World Bank (DECRG) has engaged in an extensive program of research to produce estimators The official measurement of poverty followed by the Government of Botswana is based on the poverty datum line (PDL), of welfare at geographic levels not represented in household surveys. At present, approximately 50 countries have Statistics Botswana (2013). The PDL is a Pula-denominated metric that intends to capture the cost of a basket of goods completed “poverty maps”. Moreover, in a growing number of countries, multiple-round poverty maps have been and services that would satisfy the monthly necessary and adequate requirements of a household in Botswana. The conducted or initiated to monitor the dynamics of poverty and inequality over time. The production of these maps PDL basket captures the following basic needs: food, clothing, personal items, household goods, services and shelter. is increasingly being implemented by other international organizations, partner-country statistical organizations and To account for demographic differences across households, the representative basket depends on age and gender of academic researchers. The extensive use and production of poverty maps has also benefited the assessment of this 11 the household members. To classify a household as poor, the PDL is compared to the observed total consumption of the particular small area estimation (SAE) method, as well as its extension by using other statistical approaches. household. 5 If the total consumption is below the PDL, the household is classified as poor, as well as all its members. This report is based on the SAE method developed by Elbers et al. (2000) and Elbers et al. (2003), henceforth ELL. This The national monetary value of the PDL basket in 2009/2010 was BWP878.87 of which food expenditure represented 77 method is also known as Poverty Mapping. The basic idea is to use survey data to impute welfare indicators not available percent of the total household expenditure, followed by 8 percent of other goods and 7 percent of shelter, see Statistics in census records. The main motivation is to estimate these indicators at geographical partitions not representative in Botswana (2011).6 household surveys. The suggested stages by ELL are: (a) comparability between census and survey variables; (b) modelling of income or consumption using the survey; and c) computing welfare indicators on census records (such as head count 2.2 Data source ratio, inequality) based on the parameters derived from the household survey.

The data source for the official measurement of poverty at the district and national levels is the Botswana Core Welfare 3.1 Stage zero. Comparability of Data Indicator Survey (BCWIS). The BCWIS was created as an improvement to the Household Income and Expenditures Survey (HIES) for the computation of the PDL, among other objectives. The BCWIS has been scheduled to be conducted every Prior to constructing the econometric model of expenditure or income, comparable covariates between census and five years. survey have to be identified. To do so, one identifies the geographical partition for which this comparison has to be done. This partition is country specific and is determined by the representativity of the household survey. For instance, if the The BCWIS uses a two-stage stratified probability sample design, in which enumeration areas (EA) are chosen as primary survey is representative at the region, province or district level, several models can be estimated at these geographical sampling units in the first stage, and occupied households within each EA are chosen in the second stage. These occupied partitions. However, the lower the level of disaggregation of the model, the smaller the number of observations for households form the basis of secondary sampling units. 8 estimating parameters, which in turn reduces the predictive power of the model. In our case, the team has decided to use a national model and has included interactions of district and region dummies with other covariates. To strengthen the accuracy of the information, household consumption data were collected over a period of one month, covering all household members and visitors that spent the night before interviewing the household, as long as they were After defining the geographical partition for modelling, one proceeds with the selection of comparable variables between expected to stay in the same household for at least 14 days. Prisons, hospitals, army barracks, hotels, camps, commercial census and survey. To define a set of variables strictly comparable, definitions of each variable should be compared farms and other institutional dwellings were not included in the survey. Households in completely industrial areas were also between census and survey based on their questionnaires. Subsequently, the comparison of distributions and statistics out of the scope of the survey, as well as non-citizen tourists who were in Botswana on holiday (not working). determines the final set of variables to be used for the modelling stage. Statistical comparisons between census and survey are carried out controlling for the survey sample design. For instance, standard errors must consider the clustering 2.3 An Overview of Botswana’s Poverty of primary sample units, or other stratification used as part of the sample design.

According to the World Bank (2014), Botswana is considered an upper middle income country with a GNI per capita 3.2 Stage one. Expenditure modelling (Atlas method, current US$) of $7,730; this is equivalent to 4.8 times above the average for Sub-Saharan Africa (SSA). Based on the BCWIS, 19.3 percent of the population live below the poverty line or PDL.9 Once comparable variables between census and survey are identified, one models income or consumption using the survey. The basic idea is to first estimate an Ordinary Least Squares (OLS) model with the set of comparable variables. The Although Botswana has faced a decrease in poverty incidence, going from 30.6 percent in 2003 to 19.3 percent in empirical model follows this structure: 2009/2010, in certain areas of the country poverty remains extremely high (e.g. Ngamiland West (46.2 percent), Ngwaketse West (41.7 percent), Central Bobonong (32.8 percent) and Kweneng West (32.4 percent)). Nonetheless, the districts with lnY_(ch)=X(ch) β+U(ch) (1) the highest absolute number of poor people are Kweneng East (45,557 people), Central Serowe/Palapye (43,076 people), Central Tutume (28,735 people) and Central Bobonong (25,385 people). These four districts house around 40 percent of where lnYch) denotes the logarithm of the household per capita adult-equivalent expenditure h belonging to the cluster c;

the total poor population in Botswana. X(ch) represents household and dwelling characteristics ; and U(ch) corresponds to the error component. The latter may be decomposed into cluster and household effect: Consistent with results seen in other countries, cities and towns tend to show a lower poverty incidence than rural areas. In E this case, poverty in cities and towns is at 8 percent (10.6 percent in 2002/03) and 24.3 percent (44.8 percent in 2002/03) U(ch)=ηc+ (ch) in rural areas. . In urban villages, the poverty rate is 19.9 percent (25.4 percent in 2002/03). 10

At the district level, poverty incidence decreased in all districts except for Sowa Town is 6.8 percent (3.4 percent in 2002/03). However, the number of poor people is 240 (93 in 2002/2003), representing 0.07 percent of the total number of the poor population in the country. On the other hand, 4 of the 26 surveyed districts present poverty rates above 30 percent (16 districts in 2002/03), with Ngamiland West ranking as the poorest district with 46.2 percent of people living below the poverty line. The most populated districts, housing 35.8 percent of Batswana, have poverty rates of 6.1 percent (Gaborone), 17.6 percent (Southern/Ngwaketse District), 17.8 percent (Kweneng East) and 32.4 percent (Kweneng West).

8See Statistics Botswana (2013) for further details. 11A custom-made poverty mapping software has been produced by the World Bank in parallel with the methodology to ease the computational burden 9This subsection is based on Statistics Botswana (2013). of producing poverty maps. 10Throughout the report, we use poverty rate and poverty incidence as synonyms. 12 For this study, we also consider variables at the village level constructed from the census records,

10 Mapping Poverty in Botswana 2010 Statistics Botswana Statistics Botswana Mapping Poverty in Botswana 2010 11 E where there is neither correlation between ηc and (ch), nor between household error components. Household effect refers Household level variables such as household tenure, water source, some classifications of type of toilet, lighting, cooking to unobservable characteristics intrinsic to the family, such as ability and motivation to work of household members, and heating are comparable between both sources. However, income sources and assets are, in the majority of cases, as well as socio-economic variables not collected by the survey that may affect the level of expenditure. In addition, not comparable between survey and census. Regarding individual characteristics, some classifications of civil status, cluster effect captures unobservable features at geographical partitions above household level, such as local prices, disability, schooling and a few industry variables of the household head are comparable. We also created several heterogeneity of returns to schooling, infrastructure, among others. household size groups for which the classifications representing a large number of household members were significantly comparable between census and survey. The identification of a geographical level for representing the cluster effect is a crucial decision that may affect the standard errors of the welfare indicator projected in census data. Cluster effect defines the partition from which the For continuous variables, tests of equality of distributions have been calculated in the Poverty Map software. After main geographical variability exists. The empirical distribution of this partition is used for obtaining a similar vector in comparing cumulative distributions, the following continuous variables were comparable between census and survey: census records. Hence, the lower the level of disaggregation the better it is captured the heterogeneity at the target age, household size, proportion of children between 0-15, 12-17, and 6-11, proportion of females, proportion of adults geographical level. However, this comes at the cost of having a less reliable distribution derived from survey data. ELL between 18-64 and older than 65, proportion of household members that are currently working, proportion of adults, (2003) suggests the definition of the cluster effect at the primary sample unit (PSU) or enumeration area level. However, number of rooms in the dwelling and proportion of adults that attended school. Using the comparable variables, several recent empirical literature has shown that defining the cluster effect at such level may provide “naive” standard errors of econometric models were estimated. The model with the highest root mean square error (RMSE) and adjusted R2 was poverty estimates. This recent evidence suggests the definition of the cluster effect at the level of the target indicator. The used to estimate poverty rates at the village level. It is worth mentioning that the limited number of comparable variables current Poverty Map defines the cluster effect at the village level for village poverty estimates and at the district level for between census and survey represents a concern for reducing the household effect in our econometric models. The district poverty estimates. main consequence of this limitation is the low prediction power of econometric models and as a result, this will be reflected on the precision of Poverty Map estimates. The following section discusses the model selection, main findings After defining the cluster effect, the ELL method suggests a model of the household error component, assuming a and the precision of our Poverty Map estimates. heteroskedastic structure. Subsequently, household residuals, netted out from the cluster effect, are used to correct the standard errors through Generalized Least Square (GLS) estimation. 5. Results

3.3 Stage two. Computation of the welfare measure Several econometric models of per capita adult-equivalent consumption were estimated by using the set of comparable variables, interactions and village level variables constructed with census records.14 The next sections summarize the Once the econometric model of equation (1) is estimated, its coefficients and error components are used as a bridge model selection and the main findings of the Poverty Map. We suggest the reader to skip section 5.1 if he is not interested to obtain welfare indicators in census records through the variables used in the model. Because coefficients and errors in the details of the different models. present a mean and a standard error, these are used to simulate several replications of these parameters. As a result, the ELL method produces several replications of income or expenditure at the household level that are contrasted with 5.1 Model Selection the poverty line to determine if a household lives in poverty or not. Once these households are identified, these are aggregated at the target geographical level. The current study considered a national level model including several interactions of district and regional dummies with comparable variables. We modeled the logarithm of adult equivalent consumption (LNEXP) and the logarithm of 4. Data the per capita adult equivalent consumption (per capita LNEXP). The RMSEs of the LNEXP models range from 0.7590 to 0.7728 and the RMSEs of the per capita LNEXP range from 0.749 to 0.805. The adjusted R2s present values from 0.27 to 0.30 The Poverty Map of Botswana uses the Botswana Core Welfare Indicators Survey (BCWIS 2009/2010) and the Population for the LNEXP models and from 0.48 to 0.54 for the per capita models. We decided to carry out the rest of the analysis Census 2011. Using the survey weights, the total number of population represented by 27,222 sampled observations is based on per capita LNEXP models. 1,861,548 inhabitants in Botswana. The census data contains 2,024,904 individuals for which 1,949,142 live in non-institutional households. Individuals living in institutional households and belonging to localities with no affiliation were not considered Table A presents the results of five econometric models of per capita LNEXP. Model 1 includes only comparable variables, in our Poverty Map (75,762 individuals living in prisons, orphanages, among other institutions and 19,520 living in localities Model 2 includes comparable variables and interactions with district and region dummies, Model 3 includes comparable with no affiliation). The final number of observations used from the census data, accounting for the missing values in model variables, interactions and village level variables, and Model 4 and 5 are more parsimonious models (fewer variables) covariates was approximately 1.8 million people. using all types of variables as in Model 3. Based on the highest adjusted R2, the smallest RMSE and the smallest cluster effect at the village level, Model 3 was identified as the best model for carrying out the Poverty Mapping. The poverty line used for the Poverty Mapping exercise was the mean of the PDL at the regional level using the survey. The poverty lines expressed in pulas are: 1,371 pulas for Region 1; 1,478 for Region 2; 1,265 for Region 3; 1,577 for Region Table A: Model Selection Per Capita Logarithm of Adult Equivalent 4; 1,397 for Region 5; 1,347 for Region 6; 1,321 for Region 7; 1,169 for Region 8; and 1,332 for Region 9. Poverty lines were 2 divided by household size given that the consumption aggregate is expressed in per capita terms. Adj R RMSE Cluster Effect No. Variables Model 1 0.5 0.782 0.049 28 Tables 1 to 6 present the summary statistics of census and survey variables that present comparable definitions in both Model 2 0.52 0.762 0.045 38 questionnaires. Approximately 200 variables were created for which 40 percent of them is comparable between census Model 3 0.54 0.749 0.027 63 and survey. The comparability is higher for household than for individual variables. For the latter, the percentage of Model 4 0.53 0.758 0.041 46 variables statistically comparable is 23.1 percent (at the 90 percent confidence level) and 40.4 percent (at the 99 percent confidence level). In contrast, for household variables the range falls between 40.4 to 52.2 percent for the same confidence Model 5 0.48 0.805 0.05 18 levels. The current exercise is based on the selection of variables using the 95 percent confidence level criterion. Note: Model 1 uses only comparable variables, Model 2 uses comparable variables and interactions, and Models 3 to 5 contain comparable variables, interactions and village lable variables

13The regions are compounded by the following districts: Region 1 Selibe Phikwe, Orapa, Sowa Town, Central Serowe/Palapye, Central Mahalapye, Central Bobonong, Central Boteti, and Central Tutume; Region 2 Ngamiland East, Ngamiland West, and Chobe; Region 3 Ghanzi; Region 4 Kgalagadi 14The Statistics Botswana and the Botswana’s Poverty Eradication Team kindly provided school and health data at the village level for most of the South and Kgalagadi North, Region 5 Kgatleng; Region 6 Kweneng East and Kweneng West; Region 7 Francistown and North East; Region 8 Gaborone, villages in Botswana. The lack of information of some villages did not allow us to use these variables in our econometric models. Lobatse, and South East; and Region 9 Jwaneng, Southern, Barolong, and Ngwaketse West. 12 Mapping Poverty in Botswana 2010 Statistics Botswana Statistics Botswana Mapping Poverty in Botswana 2010 13 Figure 1 of Appendix of Figures present residual diagnostic plots for Models 1 to 3. Model 1 and 2 present some extreme From the twenty villages with the highest poverty incidence, the districts containing the poorest villages are Ngwaketse outliers that may be detrimental for our final poverty estimation. Model 3 does not indicate extreme outliers with exception and Ngwaketse West. The twenty poorest villages present poverty rates from 62 percent to 77 percent. These villages of 2 values close to 4 and -4. Figure 2 shows the residuals of the consumption models 1 2 and 3 (these models are also concentrate 3.5 percent of the poorest population of the country. known as beta models based on the ELL method). We observe that the tales of the distributions gradually shrink and the shape of the distributions appear more symmetric. To summarize the geographical dispersion of village poverty incidence, Figures 5 and 6 present the visual representation of Botswana showing the poverty incidence and density at the village level, respectively. 18 Figure 5 shows the proportion of Finally Figures 3 and 4 present quartile and quartile plots. These are probability plots that show the quartiles of the residual individuals whose adult-equivalent consumption is below the poverty line to the total number of inhabitants of the village. probability distribution. The 45 degree line indicates the similarity of both residuals. In the first case, we contrast the residuals This figure reveals that the poverty incidence is concentrated in the north of the Kgalagadi district, the district borders of LNEXP with the residuals of per capital LNEXP. Figure 3 does not indicate differences between them with exception of of Kweneng and Southern/Ngwaketse districts and the north of Ngamiland. This figure also highlights the low poverty the extremes of the distribution (values below -3 and values above 3). Similarly, Figure 4 compares the residuals of Model rates around Gaborone and Francistown (see the blue circles representing poverty rates under 20 percent). However, 1 and Model 3 (per capital LNEXP models). This figure does not reveal differences between both types of residuals with this measure does not take into account the number of inhabitants in the villages. For this reason, Figure 6 presents the exception of the extreme values of the residual distributions. total number of poor individuals that exist in each village. As it is expected, those districts with the highest population density present the highest levels of poverty density. Based on this indicator, the districts with the highest concentration The model used for the Poverty Mapping (Model 3) is presented in Table 7 of Appendix of Tables, including the alpha of poor population are: the south of the North East district, south and east of the Central district and the district borders model in Table 8 and the GLS model in Table 9. The main model uses household, interactions and village level variables. of Kweneng and Southern districts. As noted before, Kweneng and Southern districts also present high poverty incidence. After running the alpha models of Model 1 to Model 5 using district dummies and stepwise selection separately, the final alpha model uses only district dummies. 15 5.2.3 Precision of poverty estimates

5.2 Main findings The precision of poverty estimates in Poverty Mapping relies on the set of comparable variables between census and survey, the prediction power of the econometric model, the number of observations in the target population and the Table 10 presents the poverty incidence at the national and regional level calculated in the survey and estimated by the type of simulation used for obtaining consumption aggregates for each household record. Even though around 40 Poverty Map. The national poverty estimated by the Poverty Map method is 22 percent. Both sources show that Region 1 percent of the constructed variables are comparable between census and survey, the precision of village level estimates and 2 are among the regions with the highest poverty incidence.16 The Poverty Map also highlights Region 9 as a region is limited. For this reason, we provide a further discussion about the precision of poverty estimates based on their standard with high poverty incidence. By looking at the confidence intervals of Region 2 and Region 9, we observe that these errors derived from the Poverty Mapping exercise. regions present much higher confidence intervals than the national poverty rate. Although the comparison between the survey and poverty map estimates is a “naive” assessment of our main findings, Table 10 shows the ranking of poverty Figure 7 displays village poverty rates and their corresponding 90 and 95 percent confidence intervals. In general, these estimates at the regional level using the BCWIS and the Poverty Map estimates including their confidence intervals. confidence intervals look uniform across the village poverty distribution. This means that the precision of high and low poverty rates is similar. However, village poverty rates between 30 to 40 percent are highly imprecise. In addition, these 5.2.1 District Poverty in Botswana plots also help in identifying the statistically differences among villages. For instance, poverty rates between 50 to 70 percent are significantly different from the villages between 0 to 25 percent. Because the confidence interval of poverty The focus of this report is primarily on village poverty rates. However, we provide to the reader the district poverty rates rates between 50 to 70 percent falls inside the confidence interval of poverty rates between 25 to 50 percent, we cannot estimated by the Poverty Map given that these are contrasted with village poverty rates in the following subsection. To conclude that there are significant differences among village poverty rates between 25 to 70 percent at the 90 percent obtain these estimates, the cluster effect was defined at the district level. confidence level.

When survey and Poverty Map district poverty rates are contrasted, in the majority of cases the Poverty Map produces To complement the discussion of the precision of village poverty estimates, we provide pairwise comparisons. These more precise poverty rates than the ones calculated in the survey. Some of the districts that present better precision in comparisons help in understanding how different the poverty incidence of village A is from village B within the country or their poverty rates are: Lobatse, Jwaneng, Chobe, Ghanzi and Kgalagadi South. As shown in Table 11, the majority of within the district. Figure 8 of the Appendix of Figures displays the percentage of pairwise comparisons among all villages district poverty estimates present coefficients of variation under 20 percent calculated by the Poverty Map, whereas for in the country. This figure shows the percentage of villages that present statistically significant differences among them. about 50 percent of districts in the survey present coefficients of variation bigger than 20 percent. 17 For instance, 35 percent of the pairwise comparisons among village poverty estimates are significantly different at the 95 percent confidence level. However, after applying a Bonferroni correction (significance level over the total number District poverty estimates show Ngwaketse West (50 percent), Ngamiland West (38 percent), Okavango Delta (35 of pairwise comparisons), we observe in Figure 9 that around 4 percent of the comparisons are significantly different at percent), Central Bobonong (35 percent) and Kweneng West (35 percent) as the poorest districts in Botswana. These the 95 percent confidence level. Because this percentage is smaller than the chosen significance level, based on the districts are followed by Ngwaketse (33 percent), Barolong (33 percent), Central Boteti (33 percent) and Ngamiland East Bonferroni correction we cannot conclude that village poverty rates are significantly different within the country. It is worth (31 percent). highlighting that the number of village pairwise comparisons is 124,750. To find significant differences among villages at the 95 percent confidence level our village poverty rates must be extremely precise. In other words, the p-value of our 5.2.2 Village poverty in Botswana pairwise comparisons must be lower than 0.0000004 (alpha/total number of pairwise comparisons = 0.05/124,750). For this reason, we provide to the reader the uncorrected and Bonferroni corrected pairwise comparisons.19 Table 12 summarizes the twenty villages with the lowest and highest poverty incidence. From the set of villages with the lowest poverty incidence, Gaborone, South East, Francistown and Selibe Phikwe contain villages with low poverty rates, Figure 10 shows the percentage of statistically significant different pairwise comparisons of villages within a district at but high concentration of the poor. These villages concentrate 5.3 percent of the total number of poor population in varying confidence levels. Because there are 7 districts of 27 with only one village, these districts are not considered Botswana. It is worth highlighting that several villages enlisted in Table 12 present large standard errors as a result of their in Figure 10. The panel of the 90 percent confidence level reveals that 5 districts of 20 have villages with poverty rates small number of observations. significantly different from the villages of the district. The percentage of villages is above 15 percent. The panel of the 99 percent confidence level reveals that in 3 districts about 20 percent of the village pairwise comparisons are significantly different. This figure suggests that there are a few villages that are significantly different within the districts.

15Previous explorations with Hungarian data point out that the selection of a simple alpha model (e.g. using district or regional dummies) produces 18 Poverty density is defined as the ratio of poor population in the village to the total population living in poverty at the national level. For instance, if village smaller standard errors of Poverty Map estimates. This result is also found with the Botswana’s survey. A presents a density of 0.02, this is interpreted as 2 percent of the total number of poor population is concentrated in the village A. Figure 4 is represented by groups of poor population, the total number of poor individuals in the country, based on the Poverty Map estimates, is 376,913 people. 16Poverty incidence corresponds to the total number of people that belongs to a specific geographical partition that presents an adult equivalent consumption below the poverty line over the total number of people living/belonging to that geographical partition. For regional poverty, the poverty 19The Bonferroni Correction aims at minimizing the Type I error (rejection of the null hypothesis when it is true). However, this correction comes at the cost of incidence corresponds to the ratio of poor population belonging to or living in a specific region to the total population of the region. increasing the Type II error (non-rejection of the null hypothesis when it is false). The team provides the uncorrected and corrected pairwise comparisons in Figures 8 and 9 to provide a more complete picture of the precision of village poverty estimates. The rest of pairwise comparisons use Bonferroni 17 The coefficient of variation is the ratio of the poverty standard error to the poverty rate. corrections 14 Mapping Poverty in Botswana 2010 Statistics Botswana Statistics Botswana Mapping Poverty in Botswana 2010 15 What is the application of the main findings for policy making? We must recall the main purpose of our Poverty Map. References For the Botswana’s case, the main aim of this exercise is to understand the heterogeneity of districts by analysing village poverty incidence. The advantage of understanding how different these villages are from each other helps in efficiently M.C. Araujo, F.H.G. Ferreira, P. Lanjouw, and B. ¨ Ozler. Local inequality and project choice: Theory and evidence from allocating government resources to the poorest population. If village poverty is quite homogenous within the district, ecuador. Journal of Public Economics, 92(5):1022–1046, 2008. there is no evidence in favour of public resource allocation at the village level. T. Bedi and A. Coudouel. More than a pretty picture: using poverty maps to design To provide further evidence about how different district poverty rates are from village poverty rates, Figure 11 presents the better policies and interventions. World Bank Publications, 2007. graphics of the proportion of villages that present a significantly different poverty rate from the district poverty rate. This figure shows that at the 95 percent confidence level 8 districts of 27 have villages with significantly different poverty rates Botswana Central Statistics Office. 2011 Population & Housing Census. 2011. from the district poverty rate. In about 7 districts up to 20 percent of the villages significantly differ in their poverty rate from the district one. In one district, 45 percent of its villages significantly differ from the district poverty rate. C. Elbers, J.O. Lanjouw, and P. Lanjouw. Welfare in villages and towns: Microestimation of poverty and inequality. Technical report, Discussion Paper TI 2000-029/2, Tinbergen Institute, Amsterdam (http:/www. tinbergen. nl), 2000. 5.2.4 Further checks: Application of Empirical Bayes Prediction C. Elbers, J.O. Lanjouw, and P. Lanjouw. Micro–level Estimation of Poverty and Inequality. Econometrica, 71(1):355–364, Using a household survey and census records, the current study uses the standard ELL method to estimate village poverty 2003. ISSN 1468-0262. rates. Recent extensions of the standard ELL are based on the use of Empirical Bayes (EB) prediction to accurately predict nonlinear functions of the dependent variable (Elbers and van der Weide, 2014). The authors point out that the benefits C. Elbers, P.F. Lanjouw, J.A. Mistiaen, B. Ozler, and K. Simler. On the unequal inequality of poor communities. The World of using EB estimation might be modest when the number of domains covered by the household survey is small (between Bank Economic Review, 18(3):401–421, 2004. 5 to 25 percent of all domains in the population). However, the EB approach may have clear benefits when the number of domains is larger (between 50 to 100 percent). 20 C. Elbers and van der Weide R. “Estimation of normal mixtures in a nested error model with an application to small area estimation of poverty and inequality. Policy Research Working Paper, WPS6962. July 2014. The BCWIS covers 152 villages out of 500 (30 percent) which makes us believe that the EB approach may have a modest contribution to our Poverty Map. However, the team decided to re-run the beta model using the EB prediction. Figures 12 J. Hentschel, J.O. Lanjouw, P. Lanjouw, and J. Poggi. Combining census and survey data to trace the spatial dimensions and 13 summarize the main results. Figure 12 displays village poverty rates including their 90 and 95 percent confidence of poverty: A case study of Ecuador. The World Bank Economic Review, 14(1):147–165, 2000. intervals. Comparing Figure 7 with Figure 12, there are no clear differences between both approaches. L.F. López-Calva, A. Meléndez Martínez, E.G. Rascón Ramírez, L. Rodrıguez-Chamussy, and M. Szekely Pardo. El ingreso de It is worth mentioning that EB approach will provide the same poverty rates as the ELL method for those villages that do los hogares en el mapa de México. El Trimestre Economico, 75(300):843–896, 2008. not appear in survey data. However, for those villages appearing in the survey, the EB may differ from the ELL. As a result, Figure 13 shows the Poverty Map estimates of those villages that were sampled in the survey (152 villages). 21 This figure Statistics Botswana. Preliminary Results of the Botswana Core Welfare Indicators shows that the EB produces similar results to the ELL approach. (Poverty) Survey 2009/10. (2011/15), 2011.

6. Conclusions Statistics Botswana. Botswana Core Welfare Indicators Survey 2009/10. 1, 2013.

This report has provided consumption-based Poverty Map estimates at the district and village levels in Botswana using R. van der Weide. A review of Empirical Bayes prediction using sampling weights with poverty mapping in mind. November the Botswana Core Welfare Indicators Survey and the Population Census 2011. The poverty map has produced a range 6, 2014. of poverty rates including their standard errors for testing statistical differences among villages and between villages and other geographical levels. World Bank. World Development Indicators. 2014.

After analysing the precision of the poverty estimates at the village level, we conclude that the precision of village estimates allows a limited analysis of the differences among villages. However, the main findings highlight that at the 95 percent confidence level, 8 districts of 27 have villages with significantly different poverty rates from the district poverty rate. In about 7 districts up to 20 percent of the villages significantly differ in their poverty rate from the district poverty rate. Additionally, in one district, 45 percent of its villages significantly differ from the district poverty rate. This result provides evidence for targeting social programmes at the village level in those districts presenting villages with poverty rates significantly different from the district poverty rates. The Poverty Map also provides more reliable estimates of village and district poverty estimates than the survey data.

This report emphasizes the need to complement the analysis of poverty incidence with the identification of areas with a large concentration of poor people. In several cases, the rankings derived from both welfare indicators provide different conclusions as a result of the density of the poor population in villages.

To sum up, the Poverty Map estimates shed light on the geographic dispersion and concentration of poverty among villages. This information holds important value for policy-makers to prioritize the use of scarce resources in areas that need them most. To improve the prediction of consumption aggregates, we strongly recommend the availability of infrastructure data at varying sub-district levels for future Poverty Map exercises.

20For further details about the EB approach, see Elbers and van der Weide (2014) and van der Weide ( 2014) where the latter provides a practical implementation by using the latest version of the PovMap software 2.5. 21The village identifiers of Figures 12 and 13 are constructed by sorting the village poverty data from lowest to poorest. These identifiers are kept for Figure 13. The reader should have in mind that Figure 13 only contains 152 villages. 16 Mapping Poverty in Botswana 2010 Statistics Botswana Statistics Botswana Mapping Poverty in Botswana 2010 17 Appendix of Tables Table 2. Descriptive Statistics: Dwelling characteristics (type of toilet, lighting, cooking source, and heating) Variable Survey Census Comparability Mean Min CI-95 Max CI-95 Mean Table 1: Descriptive Statistics: Dwelling characteristics (rooms, house tenure, and water source) Toilet: Flush toilet 0.286 0.259 0.313 0.338 no Variable Survey Census Comparability Toilet: Ventilated improve pit (VIP) latrine 0.089 0.076 0.102 0.033 no Mean Min CI-95 Max CI-95 Mean Toilet: Pit latrine 0.385 0.359 0.411 0.422 no No. of rooms (continuous) 2.440 2.378 2.502 2.333 no Toilet: Flush communal toilet 0.007 0.003 0.010 0.001 no House tenure: Purchased 0.034 0.026 0.041 0.016 no Toilet: VIP communal 0.003 0.001 0.005 0.000 no House tenure: Rent BHC 0.015 0.010 0.020 0.011 yes Toilet: Pit latrine communal 0.032 0.026 0.039 0.006 no House tenure: Rent - Government 0.040 0.028 0.051 0.040 yes Toilet: Neighbour’s toilet 0.044 0.037 0.052 0.047 yes House tenure: Rent - Council 0.017 0.012 0.021 0.014 yes Toilet: None or dry compost 0.153 0.134 0.173 0.153 yes House tenure: Rent - Individual 0.242 0.218 0.267 0.251 yes Light: Electricity, petrol, diesel, other 0.472 0.446 0.497 0.451 yes House tenure: Rent - Company 0.023 0.016 0.030 0.020 yes Light: Solar power 0.001 0.001 0.002 0.002 yes House tenure: Rent - VDC 0.008 0.005 0.010 0.007 yes Light: Gas (LPG) 0.002 0.001 0.004 0.011 no House tenure: Free 0.083 0.069 0.097 0.086 yes Light: Bio gas 0.000 0.000 0.001 0.001 yes House tenure: Inherited 0.038 0.033 0.044 0.021 no Light: Wood 0.030 0.023 0.037 0.532 no House tenure: Self built 0.500 0.475 0.525 0.535 no Light: Paraffin 0.316 0.296 0.336 0.003 no Water: Piped Indoors 0.303 0.275 0.330 0.302 yes Light: Candle 0.175 0.159 0.190 0.001 no Water: Piped outdoors 0.411 0.384 0.438 0.401 yes Cooking: Electricity, petrol, diesel, other 0.135 0.120 0.149 0.178 no Water: Neighbour/Communal tap 0.181 0.157 0.206 0.206 yes Cooking: Solar power 0.001 0.000 0.001 0.001 yes Water: Bouser/Tanker 0.009 0.005 0.014 0.011 yes Cooking: Gas (LPG) 0.397 0.375 0.418 0.380 yes Water: Well 0.013 0.008 0.018 0.009 yes Cooking: Bio gas 0.035 0.025 0.045 0.009 no Water: Borehole 0.046 0.035 0.056 0.049 yes Cooking: Wood 0.411 0.385 0.436 0.414 yes Water: River/Stream 0.016 0.006 0.026 0.014 yes Cooking: Paraffin 0.021 0.017 0.026 0.015 no Water: Dam/Pan 0.011 0.005 0.016 0.007 yes Cooking: Cow dung 0.000 0.000 0.000 0.001 no Water: Rain water tank 0.001 0.000 0.002 0.001 yes Cooking: Coal 0.001 -0.001 0.002 0.000 yes Water: Spring water 0.001 0.000 0.002 0.000 yes Cooking: Charcoal 0.000 0.000 0.000 0.001 no Cooking: Crop waste 0.000 0.000 0.000 0.000 yes Heating: Electricity, petrol, diesel, other 0.241 0.220 0.262 0.168 no Heating: Solar power 0.001 0.000 0.002 0.001 yes Heating: Gas (LPG) or bio gas 0.047 0.039 0.055 0.011 no Heating: Wood 0.433 0.408 0.458 0.479 no Heating: Paraffin 0.014 0.010 0.018 0.003 no Heating: Cow dung 0.002 0.001 0.003 0.000 no Heating: Coal 0.002 0.000 0.004 0.001 yes Heating: Charcoal 0.000 0.000 0.001 0.002 no Heating: None 0.261 0.238 0.283 0.335 no

18 Mapping Poverty in Botswana 2010 Statistics Botswana Statistics Botswana Mapping Poverty in Botswana 2010 19 Table 3. Descriptive Statistics: Sources of income and assets Table 4. Descriptive Statistics: Gender, civil status, disability, and schooling of household head Variable Survey Census Comparability Variable Survey Census Comparability Mean Min CI-95 Max CI-95 Mean Mean Min CI-95 Max CI-95 Mean Income: agriculture 0.028 0.022 0.034 0.217 no Gender 0.541 0.525 0.557 0.524 no Income: enterprise 0.065 0.058 0.072 0.107 no Age in years 45.081 44.355 45.807 43.240 no Income: Wages 0.615 0.596 0.635 0.407 no Citizenship 2.310 1.927 2.694 3.546 no Income: Rent 0.022 0.018 0.027 0.007 no Botswana national 0.938 0.930 0.947 0.931 yes Income: Pensions 0.099 0.089 0.108 0.035 no Married 0.270 0.255 0.284 0.271 yes Income: Remittances 0.091 0.081 0.101 0.041 no Never Married 0.373 0.357 0.389 0.368 yes Income: Gov. benefits 0.062 0.054 0.071 0.015 no Civil union 0.203 0.188 0.218 0.253 no Income: Family transf. 0.015 0.011 0.019 0.006 no Separated 0.013 0.010 0.016 0.008 no Income: Livestock 0.353 0.333 0.373 0.565 no Divorced 0.022 0.019 0.026 0.019 yes Cattle 0.234 0.218 0.250 0.352 no Widowed 0.119 0.108 0.129 0.081 no Goats 0.231 0.215 0.247 0.336 no Disability of the eye 0.045 0.039 0.051 0.031 no Sheep 0.055 0.048 0.063 0.067 no Disability of the ear 0.009 0.006 0.012 0.011 yes Pigs 0.006 0.004 0.008 0.008 yes Disability of speech 0.001 0.000 0.002 0.002 yes Poultry 0.117 0.105 0.129 0.370 no Disability of a leg/legs 0.026 0.022 0.030 0.007 no Donkey/mule/horse 0.056 0.049 0.064 0.178 no Disability of an arm/arms 0.008 0.006 0.010 0.004 no Other livestock 0.009 0.006 0.012 0.005 no Inability to use the whole body 0.001 0.000 0.001 0.001 yes Van 0.119 0.108 0.130 0.152 no Intellectual impairment 0.003 0.002 0.004 0.001 no Tractor 0.011 0.008 0.013 0.020 no Ever attended school 2.186 2.168 2.205 2.156 no Car 0.164 0.149 0.179 0.199 no Edu: Still studying 0.041 0.034 0.047 0.034 yes Donkey Cart 0.121 0.110 0.133 0.118 yes Edu: Primary 0.237 0.223 0.251 0.240 yes Bicycle 0.125 0.113 0.136 0.100 no Edu: Secondary 0.316 0.299 0.332 0.320 yes Wheel Barrow 0.365 0.346 0.384 0.335 no Edu: Tertiary 0.209 0.192 0.227 0.238 no Machine 0.065 0.057 0.073 0.047 no Edu: Never 0.234 0.218 0.251 0.190 no Fridge/Freezer 0.411 0.390 0.432 0.437 no Edu: Missing 0.003 0.002 0.005 0.012 no Motor cycle 0.006 0.004 0.008 0.006 yes PC or laptop 0.109 0.096 0.122 0.167 no Radio 0.577 0.561 0.592 0.619 no TV 0.490 0.469 0.510 0.544 no Phone 0.086 0.076 0.096 0.109 no Mobile 0.827 0.814 0.841 0.898 no

20 Mapping Poverty in Botswana 2010 Statistics Botswana Statistics Botswana Mapping Poverty in Botswana 2010 21 Table 5. Descriptive Statistics: Employment and household size of household head Table 6. Descriptive Statistics: Other socio-demographic characteristics Variable Survey Census Comparability Variable Survey Census Comparability Mean Min CI-95 Max CI-95 Mean Mean Min CI-95 Max CI-95 Mean Worked 0.713 0.695 0.732 0.638 no No. of children 0 to 5 0.482 0.454 0.511 0.517 no Unemployed 0.053 0.044 0.061 0.067 no No. of children 6 to 11 0.467 0.438 0.496 0.463 yes Inactive 0.203 0.188 0.219 0.295 no No. of children 0 to 11 0.949 0.898 1.001 0.980 yes Unknown working condition 0.031 0.025 0.036 0.000 no No. of children 12 to 17 0.439 0.413 0.466 0.400 no Management 0.028 0.023 0.033 0.036 no No. of adults 18 to 64 1.889 1.850 1.929 1.977 no Skilled Labour 0.449 0.431 0.467 0.353 no No. of adults 65 or more 0.199 0.183 0.215 0.183 yes Unskilled Labour 0.259 0.244 0.274 0.231 no Female hh head 0.018 0.018 0.019 0.018 yes Agriculture 0.164 0.147 0.181 0.106 no Proportion of nationals 0.940 0.931 0.949 0.933 yes Industry 0.143 0.131 0.155 0.114 no Proportion of males 0.522 0.510 0.535 0.532 yes Services 0.428 0.408 0.449 0.418 yes Proportion of females 0.478 0.465 0.490 0.468 yes Mining 0.026 0.019 0.033 0.023 yes Proportion of children 0 to 5 0.088 0.083 0.092 0.094 no Manufacture and utilities 0.059 0.052 0.066 0.038 no Proportion of children 6 to 11 0.084 0.080 0.089 0.084 yes Construction 0.058 0.05 0.066 0.053 yes Proportion of children 0 to 11 0.172 0.164 0.180 0.178 yes Trade 0.067 0.058 0.075 0.077 no Proportion of children 12 to 17 0.088 0.083 0.093 0.080 no Transport and communications 0.021 0.017 0.025 0.022 yes Proportion of adults 18 to 64 0.670 0.657 0.682 0.681 yes Financial activities 0.011 0.008 0.014 0.009 yes Proportion of adults 65 or older 0.071 0.065 0.078 0.060 no Public administration 0.139 0.124 0.154 0.113 no Prop. of adults attended school 0.786 0.775 0.798 0.804 no Health and education 0.064 0.052 0.076 0.085 no Live births 1.289 1.234 1.345 1.722 no Other services 0.127 0.116 0.138 0.112 no Disability in the HH 0.557 0.529 0.585 0.000 no Household Size (continuous) 3.477 3.375 3.579 3.540 yes Disability of the eye in the HH 0.075 0.067 0.084 0.048 no Hh size (one person) 0.296 0.28 0.312 0.279 no Disability of the ear in the HH 0.023 0.018 0.027 0.020 yes Hh size 2-3 persons 0.307 0.296 0.318 0.319 no Disability of speech in the HH 0.005 0.003 0.007 0.011 no Hh size 4-5 persons 0.212 0.199 0.225 0.209 yes Disability of a leg/legs in the HH 0.045 0.039 0.051 0.014 no Hh size 6-7 persons 0.101 0.093 0.109 0.106 yes Disability of an arm/arms in the HH 0.014 0.011 0.017 0.008 no Hh size 8 or more persons 0.085 0.076 0.093 0.088 yes Inability to use body in the HH 0.002 0.001 0.003 0.003 yes Intellectual impairment in the HH 0.016 0.012 0.019 0.004 no

22 Mapping Poverty in Botswana 2010 Statistics Botswana Statistics Botswana Mapping Poverty in Botswana 2010 23 Table 7. OLS model of the logarithm of per capita adult equivalent expenditure Table 7. OLS model of the logarithm of per capita adult equivalent expenditure cont’d Variables Coefficient Std. Error P-Value Variables Coefficient Std. Error P-Value Intercept 7.06 0.21 0.000 Age*Services 0.00 0.00 0.002 Agriculture Income (Village) -0.30 0.18 0.097 Household Size*Household Size 0.01 0.00 0.000 Agriculture Proportion (Village) 0.85 0.19 0.000 Household Size*House tenure: Free -0.04 0.02 0.061 Cooking: Gas (LPG) 0.05 0.03 0.076 Household Size*Central Mahalapye 0.07 0.02 0.000 Cooking: Wood -0.08 0.04 0.044 District: Selebi Pikwe*Prop. adults 18 to 64 -0.01 0.00 0.000 Donkey Cart (Village) -0.53 0.23 0.022 District: Selebi Pikwe*Prop. children 0 to 11 0.00 0.00 0.049 Goats (Village) 1.10 0.22 0.000 District: Orapa*Prop. adults 18 to 64 -0.01 0.00 0.000 Household Size -0.33 0.02 0.000 District: Jwaneng*Prop. adults 18 to 64 -0.01 0.00 0.000 House tenure: Self built (Village) -0.49 0.13 0.000 District: Barolong*Prop. adults 18 to 64 0.00 0.00 0.024 House tenure: Rent BHC -0.18 0.08 0.024 District: Central Bobonong*prop. children 0 to 11 0.00 0.00 0.037 House tenure: Rent Individual -0.07 0.03 0.034 District: North East*Prop. adults 18 to 64 0.00 0.00 0.001 District: Lobatse -0.31 0.07 0.000 Region 8*Hosehold Size 0.02 0.01 0.012 District: Sowa Town -0.49 0.15 0.001 Region 9*Prop. Inactive people (Village) 1.01 0.33 0.003 District: Central Boteti -0.19 0.07 0.006 Region 9*Prop. of people without education (Village) -2.09 0.45 0.000 Income source: wages (Village) -0.48 0.18 0.008 Region 9*Prop. of unemployed people (Village) 2.51 0.93 0.007 Industry (Village) 1.10 0.34 0.001 Proportion of female*Proportion of female 0.00 0.00 0.000 Light: Electricity, petrol, diesel, other 0.19 0.03 0.000 Prop. of adults that attended school*Prop. of adults that worked 0.00 0.00 0.000 Married household head 0.46 0.09 0.000 Note: Number of observations 6819, SST=8344.0897, SSR=4557.1976, MSE=0.5605 RMSE=0.7487, F=131.1333, R2=0.5462, Motor Cycle 0.34 0.13 0.010 adjR2=0.5420. Mining 0.20 0.07 0.003 No. of rooms: 3 to 4 0.22 0.07 0.002 Pigs 0.49 0.13 0.000 Table 8. Alpha Model Primary education (hh head) -0.52 0.09 0.000 Variables CoefficienStd. Error P-Value Proportion of adults 18 to 64 0.00 0.00 0.000 Intercept -5.01 0.04 0.000 Proportion of adults older than 64 0.00 0.00 0.000 District: Ngwaketse West -0.60 0.33 0.073 Proportion of female -0.01 0.00 0.000 District: Kweneng East -0.34 0.11 0.003 Proportion of adults that attended school 0.00 0.00 0.002 District: Kgatleng -0.26 0.14 0.068 Radio (Village) 0.41 0.18 0.021 District: Central Bobonong -0.27 0.15 0.076 Secundary education (hh head) -0.55 0.08 0.000 District: North East -0.36 0.18 0.047 Sector: Services -0.14 0.07 0.055 District: Ngamiland East 0.48 0.15 0.002 Still studying (hh head) 0.13 0.06 0.019 Note: Number of observations=6819, SST=36389.2505, SSR=205.5504, MSE=5.3118 RMSE=2.3047, F=6.4495, R2=0.0056, adjR2=0.0048. Toilet: Flush toilet (Village) 0.24 0.04 0.000 Toilet: Ventilated improved pit latrine (Village) -0.75 0.28 0.008 Water: Piped indoors 0.15 0.04 0.000 Water: Barehole -0.15 0.06 0.019 Water: River/Stream -0.36 0.12 0.002 Age*Age 0.00 0.00 0.000 Age*household size 0.00 0.00 0.000 Age*District: Central Mahalapye 0.00 0.00 0.027 Age*District: Central Tutume 0.00 0.00 0.000 Age*Kgalagadi South 0.01 0.00 0.002 Age*Married household head -0.01 0.00 0.000 Age*Never Married household head 0.00 0.00 0.019 Age*No. rooms 3 to 4 0.00 0.00 0.018 Age*Primary education (hh head) 0.01 0.00 0.000 Age*Secondary education (hh head) 0.01 0.00 0.000

24 Mapping Poverty in Botswana 2010 Statistics Botswana Statistics Botswana Mapping Poverty in Botswana 2010 25 Table 9. GLS model of the logarithm of per capita adult equivalent expenditure Table 9. GLS model of the logarithm of per capita adult equivalent expenditure cont’d Variables Coefficien Std. Error P-Value Variables Coefficien Std. Error P-Value Intercept 6.95 0.19 0.000 Age*Services 0.00 0.00 0.002 Agriculture Income (Village) -0.29 0.16 0.080 Household Size*Household Size 0.01 0.00 0.000 Agriculture Proportion (Village) 0.78 0.17 0.000 Household Size*House tenure: Free -0.03 0.02 0.077 Cooking: Gas (LPG) 0.05 0.03 0.061 Household Size*Central Mahalapye 0.07 0.02 0.000 Cooking: Wood -0.08 0.04 0.021 District: Selebi Pikwe*Prop. adults 18 to 64 -0.01 0.00 0.000 Donkey Cart (Village) -0.63 0.21 0.003 District: Selebi Pikwe*Prop. children 0 to 11 0.00 0.00 0.037 Goats (Village) 1.15 0.20 0.000 District: Orapa*Prop. adults 18 to 64 -0.01 0.00 0.000 Household Size -0.32 0.02 0.000 District: Jwaneng*Prop. adults 18 to 64 -0.01 0.00 0.000 House tenure: Self built (Village) -0.42 0.12 0.001 District: Barolong*Prop. adults 18 to 64 0.00 0.00 0.017 House tenure: Rent BHC -0.18 0.07 0.015 District: Central Bobonong*prop. children 0 to 11 0.00 0.00 0.016 House tenure: Rent Individual -0.07 0.03 0.022 District: North East*Prop. adults 18 to 64 0.00 0.00 0.000 District: Lobatse -0.30 0.07 0.000 Region 8*Hosehold Size 0.02 0.01 0.005 District: Sowa Town -0.51 0.14 0.000 Region 9*Prop. Inactive people (Village) 0.87 0.30 0.004 District: Central Boteti -0.20 0.06 0.002 Region 9*Prop. of people without education (Village) -1.86 0.41 0.000 Income source: wages (Village) -0.43 0.17 0.010 Region 9*Prop. of unemployed people (Village) 0.00 0.00 0.000 Industry (Village) 1.12 0.31 0.000 Proportion of female*Proportion of female 0.00 0.00 0.000 Light: Electricity, petrol, diesel, other 0.19 0.03 0.000 Prop. of adults that attended school*Prop. of adults that worked 2.44 0.74 0.001 Married household head 0.46 0.08 0.000 Motor Cycle 0.33 0.12 0.005 Mining 0.20 0.06 0.002 Table 10. Regional Poverty Comparison between survey and Poverty Map estimates No. of rooms: 3 to 4 0.24 0.07 0.000 Survey Poverty Map Region Pigs 0.51 0.12 0.000 Poverty CIMin95 CIMax95 Poverty CIMin95 CIMax95 Primary education (hh head) -0.49 0.09 0.000 1 0.23 0.2 0.26 0.25 0.18 0.32 Proportion of adults 18 to 64 0.00 0.00 0.000 2 0.3 0.2 0.4 0.31 0.24 0.39 Proportion of adults older than 64 0.00 0.00 0.000 3 0.26 0.12 0.4 0.29 0.22 0.36 Proportion of female -0.01 0.00 0.000 4 0.2 0.13 0.26 0.23 0.16 0.29 Proportion of adults that attended school 0.00 0.00 0.001 5 0.2 0.13 0.27 0.2 0.14 0.26 Radio (Village) 0.46 0.16 0.004 6 0.2 0.15 0.26 0.24 0.18 0.3 Secundary education (hh head) -0.52 0.08 0.000 7 0.13 0.08 0.17 0.13 0.09 0.17 Sector: Services -0.12 0.07 0.059 8 0.09 0.06 0.12 0.06 0.04 0.09 Still studying (hh head) 0.14 0.05 0.006 9 0.2 0.15 0.25 0.32 0.25 0.39 Toilet: Flush toilet (Village) 0.25 0.04 0.000 National 0.19 0.18 0.21 0.22 0.19 0.25 Toilet: Ventilated improved pit latrine (Village) -0.70 0.26 0.006 Note: The regions are compounded by the following districts: Region 1 Selibe Phikwe, Orapa, Sowa Town, Water: Piped indoors 0.15 0.04 0.000 Central Serowe/Palapye, Central Mahalapye, Central Bobonong, Central Boteti, and Central Tutume; Region 2 Ngamiland East, Ngamiland West, and Chobe; Region 3 Ghanzi; Region 4 Kgalagadi South and Kgalagadi Water: Barehole -0.15 0.06 0.009 North, Region 5 Kgatleng; Region 6 Kweneng East and Kweneng West; Region 7 Francistown and North East; Water: River/Stream -0.36 0.11 0.001 Region 8 Gaborone, Lobatse, and South East; and Region 9 Jwaneng, Southern, Barolong, and Ngwaketse West. Poverty Map estimates at the regional and national level were estimated identifying the cluster effect at Age*Age 0.00 0.00 0.000 the regional level. Age*household size 0.00 0.00 0.000 Age*District: Central Mahalapye 0.00 0.00 0.016 Age*District: Central Tutume 0.00 0.00 0.000 Age*Kgalagadi South 0.01 0.00 0.001 Age*Married household head -0.01 0.00 0.000 Age*Never Married household head 0.00 0.00 0.013 Age*No. rooms 3 to 4 0.00 0.00 0.004 Age*Primary education (hh head) 0.01 0.00 0.000 Age*Secondary education (hh head) 0.01 0.00 0.000

26 Mapping Poverty in Botswana 2010 Statistics Botswana Statistics Botswana Mapping Poverty in Botswana 2010 27 Table 11. District Poverty: Comparison between survey and census poverty estimates Table 12.1 The twenty villages with the lowest and highest poverty incidence Region District Village Poverty Std. Error No. Poor Prop. Survey Poverty Map District Twenty Poor Villages With The Lowest Poverty Rates Poverty CIMin95 CIMax95 CoefVar Poverty CIMin95 CIMax95 CoefVar 1 Orapa Ora pa 0.01 0.004 62 0.000 (1) (2) (3 (4) (5) (6) (7) (8) 9 Jwaneng Jwaneng 0.03 0.008 342 0.001 Gaborone 0.06 0.04 0.08 0.20 0.03 0.02 0.05 0.26 8 Gaborone Gaborone 0.03 0.009 5,804 0.016 Francistown 0.08 0.04 0.12 0.27 0.10 0.06 0.13 0.18 1 Sowa town Sowa Town 0.03 0.017 101 0.000 Lobatse 0.15 0.05 0.25 0.33 0.14 0.09 0.19 0.19 1 Central Mahalapye Tumasera/Sel eka 0.03 0.028 11 0.000 Selibe Phikwe 0.14 0.06 0.22 0.28 0.11 0.07 0.14 0.16 8 South East Tlokweng 0.04 0.011 1,438 0.004 Orapa - - - - 0.01 0.00 0.01 0.57 1 Central Serowe Palapye Moeng 0.05 0.027 17 0.000 Jwaneng 0.03 -0.01 0.07 0.63 0.03 0.01 0.04 0.33 2 Okavango Katamga 0.05 0.068 2 0.000 Sowa Town 0.07 -0.03 0.17 0.76 0.03 0.00 0.06 0.47 9 Barolong Musi 0.06 0.079 1 0.000 Southern or Ngwaketse 0.18 0.12 0.23 0.16 0.33 0.26 0.39 0.10 7 North East Shashe Bridge 0.06 0.032 49 0.000 Barolong 0.24 0.15 0.34 0.20 0.33 0.26 0.39 0.10 6 Kweneng East Mogoditshane 0.07 0.015 3,022 0.008 Ngwaketse West 0.47 0.21 0.73 0.28 0.50 0.40 0.60 0.10 7 North East Masunga 0.07 0.019 296 0.001 South East 0.13 0.06 0.21 0.28 0.12 0.07 0.16 0.21 5 Kgatleng Matebeleng 0.08 0.028 160 0.000 Kweneng East 0.18 0.12 0.23 0.16 0.22 0.15 0.29 0.17 2 Okavango Daonara 0.09 0.115 3 0.000 Kweneng West 0.32 0.23 0.42 0.15 0.35 0.26 0.44 0.13 7 Francistown Francistown 0.1 0.019 7,909 0.021 Kgatleng 0.20 0.13 0.27 0.18 0.21 0.14 0.27 0.17 1 Selebi Pikwe Selebi Phikwe 0.11 0.02 4,630 0.012 Central Serowe/Palapye 0.27 0.21 0.34 0.13 0.27 0.21 0.33 0.11 5 Kgatleng Mabalane 0.11 0.034 78 0.000 Central Mahalapye 0.17 0.13 0.22 0.14 0.21 0.13 0.28 0.18 7 North East Mambo 0.11 0.074 18 0.000 Central Bobonong 0.33 0.23 0.42 0.15 0.35 0.26 0.43 0.12 5 Kgatleng Oodi 0.11 0.023 545 0.001 Central Boteti 0.31 0.22 0.4 0.15 0.33 0.19 0.47 0.21 7 North East Gulubane 0.12 0.048 93 0.000 Central Tutume 0.19 0.14 0.24 0.14 0.22 0.15 0.29 0.15 Twenty Villages with the Highest Poverty Rates North East 0.20 0.15 0.25 0.13 0.19 0.12 0.26 0.19 9 Ngwaketse Kutuku 0.62 0.102 100 0.000 Ngamiland East 0.23 0.15 0.3 0.17 0.31 0.24 0.37 0.11 9 Southern Gasita 0.62 0.064 1,201 0.003 Ngamiland West 0.46 0.29 0.64 0.19 0.38 0.28 0.47 0.13 9 Barolong Lejwana 0.62 0.105 298 0.001 Chobe 0.02 -0.02 0.06 0.88 0.20 0.14 0.25 0.15 9 Southern Sesung 0.62 0.068 440 0.001 Okavango Delta - - - - 0.35 0.21 0.49 0.20 9 Barolong Tswaaneng 0.63 0.066 540 0.001 Ghanzi 0.26 0.12 0.4 0.27 0.30 0.23 0.38 0.13 9 Southern Lotlhakane West 0.64 0.065 1,030 0.003 Kgalagadi South 0.17 0.07 0.27 0.29 0.23 0.16 0.30 0.15 9 Southern Betesankwe 0.64 0.088 245 0.001 Kgalagadi North 0.24 0.2 0.28 0.09 0.24 0.17 0.31 0.15 9 Southern Pitseng 0.64 0.059 644 0.002 Note: Central Kgalagadi Game Reserve has been omitted from the table given that only 46 individuals are recorded in the census for this district. 9 Southern Ralekgetho 0.65 0.077 269 0.001 3 Ghanzi Groote Laagte 0.65 0.064 579 0.002 9 Barolong Sekhutlane 0.66 0.061 488 0.001 9 Ngwaketse West Khonkhwa 0.67 0.063 343 0.001 1 Central Boteti Khwee 0.67 0.071 637 0.002 9 Ngwaketse West Mahotshwane 0.67 0.078 815 0.002 9 Southern Selokolela 0.68 0.05 1,395 0.004 3 Ghanzi Qabo 0.69 0.074 489 0.001 9 Southern Lorolwana 0.69 0.057 1,147 0.003 9 Southern Segwagwa 0.69 0.045 920 0.002 9 Southern Tlhankane 0.74 0.058 502 0.001 9 Southern Mokhomba 0.74 0.063 592 0.002 9 Ngwaketse West ltholoke 0.77 0.065 437 0.001

28 Mapping Poverty in Botswana 2010 Statistics Botswana Statistics Botswana Mapping Poverty in Botswana 2010 29 Table 12.2: Estimated Disaggregated Poverty Rates at Village Level Table 12.2: Estimated Disaggregated Poverty Rates at Village Level cont’d Census District Code Census District Village Name Number of Households Population Poverty Rate Standard Error of Poverty Rate Estimated District Poverty Standard Error of Estimated District Poverty Estimated Poor in Village Estimated Poor - National Proportion of Poor in Village Census District Code Census District Village Name Number of Households Population Poverty Rate Standard Error of Poverty Rate Estimated District Poverty Standard Error of Estimated District Poverty Estimated Poor in Village Estimated Poor - National Proportion of Poor in Village

1 Gaborone Gaborone 59,136 181,385 0.032 0.009 0.031 0.008 5,804 374,233 0.016 11 Barolong Bethel 73 316 0.155 0.064 0.326 0.032 49 374,233 0.000 2 Francistown Francistown 25,706 80,705 0.098 0.019 0.097 0.017 7,909 374,233 0.021 11 Barolong Dinatshana 112 591 0.61 0.107 0.326 0.032 361 374,233 0.001 3 Lobatse Lobatse 6,789 21,573 0.137 0.025 0.139 0.026 2,956 374,233 0.008 11 Barolong Ngwatsau 69 376 0.488 0.088 0.326 0.032 183 374,233 0.000 4 Selebi Phikwe Selebi Phikwe 13,986 43,271 0.107 0.02 0.107 0.017 4,630 374,233 0.012 11 Barolong Ramatlabama 388 1,230 0.217 0.047 0.326 0.032 267 374,233 0.001 5 Orapa Orapa 3,211 8,857 0.007 0.004 0.007 0.004 62 374,233 0.000 11 Barolong Good Hope 1,316 4,224 0.135 0.026 0.326 0.032 570 374,233 0.002 6 Jwaneng Jwaneng 4,529 13,162 0.026 0.008 0.027 0.009 342 374,233 0.001 11 Barolong Mokatako 310 1,182 0.344 0.053 0.326 0.032 407 374,233 0.001 7 Sowa Town Sowa Town 1,145 3,055 0.033 0.017 0.032 0.015 101 374,233 0.000 11 Barolong Tswanyaneng 147 604 0.272 0.06 0.326 0.032 164 374,233 0.000 10 Southern Kanye 11,828 46,887 0.236 0.03 0.327 0.034 11,065 374,233 0.030 11 Barolong Metlojane 172 721 0.305 0.058 0.326 0.032 220 374,233 0.001 10 Southern Ranaka 798 3,196 0.359 0.05 0.327 0.034 1,147 374,233 0.003 11 Barolong Borobadilepe 91 300 0.232 0.066 0.326 0.032 70 374,233 0.000 10 Southern Lotlhakane West 318 1,614 0.638 0.065 0.327 0.034 1,030 374,233 0.003 11 Barolong Hebron 164 707 0.167 0.058 0.326 0.032 118 374,233 0.000 10 Southern Gasita 547 1,947 0.617 0.064 0.327 0.034 1,201 374,233 0.003 11 Barolong Logagane 70 370 0.575 0.09 0.326 0.032 213 374,233 0.001 10 Southern Lorolwana 413 1,660 0.691 0.057 0.327 0.034 1,147 374,233 0.003 Tswagare/Lothoje/ 11 Barolong Lokalana 86 360 0.539 0.076 0.326 0.032 194 374,233 0.001 10 Southern Lehoko 117 514 0.606 0.073 0.327 0.034 311 374,233 0.001 11 Barolong Makokwe 33 132 0.309 0.123 0.326 0.032 41 374,233 0.000 10 Southern Tsonyane 202 727 0.509 0.067 0.327 0.034 370 374,233 0.001 11 Barolong Marojane 75 382 0.333 0.078 0.326 0.032 127 374,233 0.000 10 Southern Kgomokasitwa 272 1,165 0.22 0.061 0.327 0.034 256 374,233 0.001 11 Barolong Papatlo 101 477 0.304 0.077 0.326 0.032 145 374,233 0.000 10 Southern Pitseng 226 1,003 0.642 0.059 0.327 0.034 644 374,233 0.002 11 Barolong Phihitshwane 120 648 0.221 0.06 0.326 0.032 143 374,233 0.000 10 Southern Mokhomba 193 799 0.741 0.063 0.327 0.034 592 374,233 0.002 11 Barolong Molete 59 305 0.565 0.091 0.326 0.032 172 374,233 0.000 10 Southern Lekgolobotlo 202 920 0.345 0.069 0.327 0.034 317 374,233 0.001 11 Barolong Ditlharapa 133 550 0.382 0.076 0.326 0.032 210 374,233 0.001 10 Southern Seherelela 130 653 0.609 0.073 0.327 0.034 398 374,233 0.001 11 Barolong Madingwana 83 282 0.262 0.07 0.326 0.032 74 374,233 0.000 10 Southern Lotlhakane 997 4,486 0.297 0.04 0.327 0.034 1,332 374,233 0.004 11 Barolong Kgoro 149 614 0.459 0.078 0.326 0.032 282 374,233 0.001 10 Southern Sese 617 2,603 0.319 0.056 0.327 0.034 830 374,233 0.002 11 Barolong Sheep Farm 53 274 0.543 0.084 0.326 0.032 149 374,233 0.000 10 Southern Sesung 146 707 0.622 0.068 0.327 0.034 440 374,233 0.001 11 Barolong Motsentshe 95 393 0.488 0.081 0.326 0.032 192 374,233 0.001 10 Southern Magotlhwane 265 1,316 0.408 0.055 0.327 0.034 537 374,233 0.001 11 Barolong Mogwalale 81 325 0.353 0.076 0.326 0.032 115 374,233 0.000 10 Southern Segwagwa 299 1,327 0.693 0.045 0.327 0.034 920 374,233 0.002 11 Barolong Gathwane 281 997 0.193 0.057 0.326 0.032 192 374,233 0.001 10 Southern Manyana 793 3,403 0.295 0.047 0.327 0.034 1,004 374,233 0.003 11 Barolong Digawana 678 2,847 0.179 0.037 0.326 0.032 510 374,233 0.001 10 Southern Maokane 479 1,715 0.54 0.057 0.327 0.034 926 374,233 0.002 11 Barolong Magoriapitse 277 961 0.351 0.056 0.326 0.032 337 374,233 0.001 10 Southern Dipotsana 44 120 0.183 0.093 0.327 0.034 22 374,233 0.000 11 Barolong Lejwana 92 480 0.62 0.105 0.326 0.032 298 374,233 0.001 10 Southern Diabo 91 232 0.255 0.083 0.327 0.034 59 374,233 0.000 11 Barolong Mogojogojo 394 1,756 0.396 0.065 0.326 0.032 695 374,233 0.002 10 Southern Molapowabojang 2,136 8,719 0.212 0.037 0.327 0.034 1,848 374,233 0.005 11 Barolong Mmathethe 1,751 6,075 0.393 0.045 0.326 0.032 2,387 374,233 0.006 10 Southern Ralekgetho 93 417 0.645 0.077 0.327 0.034 269 374,233 0.001 11 Barolong Mokgomane 235 721 0.361 0.062 0.326 0.032 260 374,233 0.001 10 Southern Moshaneng 347 1,627 0.607 0.059 0.327 0.034 988 374,233 0.003 11 Barolong Pitshane Molopo 567 1,702 0.193 0.04 0.326 0.032 328 374,233 0.001 10 Southern Moshupa 5,166 20,935 0.317 0.035 0.327 0.034 6,636 374,233 0.018 11 Barolong Sedibeng 150 703 0.359 0.066 0.326 0.032 252 374,233 0.001 10 Southern Ntlhantlhe 434 1,891 0.395 0.051 0.327 0.034 747 374,233 0.002 11 Barolong Musi 16 20 0.057 0.079 0.326 0.032 1 374,233 0.000 10 Southern Tshwaane 40 144 0.387 0.111 0.327 0.034 56 374,233 0.000 11 Barolong Tswaaneng 212 854 0.632 0.066 0.326 0.032 540 374,233 0.001 10 Southern Semane 264 720 0.482 0.068 0.327 0.034 347 374,233 0.001 11 Barolong Gamajalela 36 93 0.216 0.09 0.326 0.032 20 374,233 0.000 10 Southern Tlhankane 126 679 0.739 0.058 0.327 0.034 502 374,233 0.001 11 Barolong Dikhukhung 70 269 0.405 0.082 0.326 0.032 109 374,233 0.000 10 Southern Selokolela 475 2,060 0.677 0.05 0.327 0.034 1,395 374,233 0.004 11 Barolong Leporung 170 639 0.462 0.079 0.326 0.032 295 374,233 0.001 10 Southern Mogonye 117 521 0.255 0.064 0.327 0.034 133 374,233 0.000 11 Barolong Mmakgori 172 760 0.495 0.071 0.326 0.032 376 374,233 0.001 10 Southern Betesankwe 87 383 0.639 0.088 0.327 0.034 245 374,233 0.001 11 Barolong Mabule 387 1,568 0.474 0.059 0.326 0.032 743 374,233 0.002 11 Barolong Pitsane Siding 667 2,680 0.163 0.032 0.326 0.032 437 374,233 0.001 11 Barolong Tshidilamolomo 196 716 0.208 0.061 0.326 0.032 149 374,233 0.000 11 Barolong Tlhareseleele 126 540 0.354 0.061 0.326 0.032 191 374,233 0.001 11 Barolong Metlobo 333 1,086 0.462 0.063 0.326 0.032 502 374,233 0.001 11 Barolong Pitsana-Potokwe 185 797 0.219 0.054 0.326 0.032 175 374,233 0.000 11 Barolong Lorwana 265 1,113 0.318 0.053 0.326 0.032 354 374,233 0.001 11 Barolong Rakhuna 195 838 0.231 0.048 0.326 0.032 194 374,233 0.001 11 Barolong Kangwe 62 237 0.516 0.088 0.326 0.032 122 374,233 0.000 11 Barolong Malokaganyane 43 196 0.495 0.093 0.326 0.032 97 374,233 0.000

30 Mapping Poverty in Botswana 2010 Statistics Botswana Statistics Botswana Mapping Poverty in Botswana 2010 31 Table 12.2: Estimated Disaggregated Poverty Rates at Village Level cont’d Table 12.2: Estimated Disaggregated Poverty Rates at Village Level cont’d Census District Code Census District Village Name Number of Households Population Poverty Rate Standard Error of Poverty Rate Estimated District Poverty Standard Error of Estimated District Poverty Estimated Poor in Village Estimated Poor - National Proportion of Poor in Village Census District Code Census District Village Name Number of Households Population Poverty Rate Standard Error of Poverty Rate Estimated District Poverty Standard Error of Estimated District Poverty Estimated Poor in Village Estimated Poor - National Proportion of Poor in Village 30 Kweneng East Kweneng 163 650 0.45 0.079 0.218 0.036 293 374,233 0.001 11 Barolong Sekhutlane 166 736 0.663 0.061 0.326 0.032 488 374,233 0.001 30 Kweneng East Ditshukudu 52 178 0.217 0.082 0.218 0.036 39 374,233 0.000 12 Ngwaketse West Mabutsane 695 2,062 0.157 0.044 0.496 0.051 324 374,233 0.001 30 Kweneng East Mokolodi 75 515 0.416 0.083 0.218 0.036 214 374,233 0.001 12 Ngwaketse West Morwamosu 203 729 0.383 0.085 0.496 0.051 279 374,233 0.001 30 Kweneng East Mmatseta 120 482 0.177 0.05 0.218 0.036 85 374,233 0.000 12 Ngwaketse West Sekoma 366 1,425 0.448 0.064 0.496 0.051 638 374,233 0.002 30 Kweneng East Leologane 158 805 0.582 0.097 0.218 0.036 469 374,233 0.001 12 Ngwaketse West Khonkhwa 120 514 0.668 0.063 0.496 0.051 343 374,233 0.001 30 Kweneng East Gakutlo 400 1,656 0.272 0.057 0.218 0.036 450 374,233 0.001 12 Ngwaketse West Keng 283 1,115 0.518 0.064 0.496 0.051 578 374,233 0.002 30 Kweneng East Dikgatlhong 61 204 0.253 0.091 0.218 0.036 52 374,233 0.000 12 Ngwaketse West Khakhea 691 3,038 0.605 0.059 0.496 0.051 1,838 374,233 0.005 30 Kweneng East Medie 95 416 0.445 0.086 0.218 0.036 185 374,233 0.000 12 Ngwaketse West Kokong 288 1,069 0.483 0.072 0.496 0.051 516 374,233 0.001 31 Kweneng West Letlhakeng 2,335 8,985 0.307 0.042 0.349 0.044 2,758 374,233 0.007 12 Ngwaketse West Kanaku 64 216 0.377 0.15 0.496 0.051 81 374,233 0.000 31 Kweneng West Botlhapatlou 514 1,822 0.306 0.052 0.349 0.044 558 374,233 0.001 12 Ngwaketse West Mahotshwane 302 1,209 0.674 0.078 0.496 0.051 815 374,233 0.002 31 Kweneng West Ditshegwane 549 2,409 0.397 0.049 0.349 0.044 956 374,233 0.003 12 Ngwaketse West Itholoke 118 568 0.769 0.065 0.496 0.051 437 374,233 0.001 31 Kweneng West Dutlwe 339 1,266 0.325 0.054 0.349 0.044 411 374,233 0.001 12 Ngwaketse West Kutuku 36 162 0.616 0.102 0.496 0.051 100 374,233 0.000 31 Kweneng West Moshaweng 807 2,675 0.199 0.048 0.349 0.044 532 374,233 0.001 20 South East Otse 1,489 5,198 0.152 0.034 0.115 0.024 790 374,233 0.002 31 Kweneng West Khudumelapye 767 2,694 0.334 0.052 0.349 0.044 900 374,233 0.002 Ramotswa Station/ 20 South East Taung 1,111 3,951 0.133 0.033 0.115 0.024 525 374,233 0.001 31 Kweneng West Serinane 150 584 0.405 0.06 0.349 0.044 237 374,233 0.001 20 South East Ramotswa 6,430 27,632 0.183 0.034 0.115 0.024 5,057 374,233 0.014 31 Kweneng West Mantshwabisi 169 646 0.274 0.071 0.349 0.044 177 374,233 0.000 20 South East Mogobane 711 2,558 0.244 0.045 0.115 0.024 624 374,233 0.002 31 Kweneng West Motokwe 536 1,945 0.28 0.059 0.349 0.044 545 374,233 0.001 20 South East Tlokweng 10,969 32,674 0.044 0.011 0.115 0.024 1,438 374,233 0.004 31 Kweneng West Ngware 354 1,243 0.334 0.06 0.349 0.044 415 374,233 0.001 30 Kweneng East Molepolole 15,073 65,406 0.297 0.048 0.218 0.036 19,426 374,233 0.052 31 Kweneng West Salajwe 617 2,398 0.318 0.051 0.349 0.044 763 374,233 0.002 30 Kweneng East Lentsweletau 1,998 7,010 0.26 0.051 0.218 0.036 1,823 374,233 0.005 31 Kweneng West Takatokwane 777 2,906 0.354 0.055 0.349 0.044 1,029 374,233 0.003 30 Kweneng East Mahetlwe 266 693 0.116 0.044 0.218 0.036 80 374,233 0.000 31 Kweneng West Tsetseng 98 438 0.312 0.074 0.349 0.044 137 374,233 0.000 30 Kweneng East Gakgatla 159 696 0.205 0.063 0.218 0.036 143 374,233 0.000 31 Kweneng West Tswaane 133 593 0.559 0.086 0.349 0.044 331 374,233 0.001 30 Kweneng East Gamodubu 118 541 0.278 0.072 0.218 0.036 150 374,233 0.000 31 Kweneng West Monwane 98 491 0.552 0.084 0.349 0.044 271 374,233 0.001 30 Kweneng East Gabane 4,038 14,789 0.129 0.026 0.218 0.036 1,908 374,233 0.005 31 Kweneng West Malwelwe 220 1,054 0.381 0.059 0.349 0.044 402 374,233 0.001 30 Kweneng East Hatsalatladi 452 1,436 0.277 0.06 0.218 0.036 398 374,233 0.001 31 Kweneng West Maboane 506 1,841 0.34 0.057 0.349 0.044 626 374,233 0.002 30 Kweneng East Ramaphatle 108 677 0.502 0.073 0.218 0.036 340 374,233 0.001 31 Kweneng West Sesung 319 1,373 0.378 0.065 0.349 0.044 519 374,233 0.001 30 Kweneng East Kopong 2,145 8,954 0.203 0.036 0.218 0.036 1,818 374,233 0.005 31 Kweneng West Sorilatholo 162 748 0.544 0.073 0.349 0.044 407 374,233 0.001 30 Kweneng East Tloaneng 147 711 0.274 0.071 0.218 0.036 195 374,233 0.001 31 Kweneng West Kotolaname 105 529 0.516 0.077 0.349 0.044 273 374,233 0.001 30 Kweneng East Kumakwane 1,343 5,341 0.229 0.038 0.218 0.036 1,223 374,233 0.003 31 Kweneng West Metsibotlhoko 91 372 0.305 0.078 0.349 0.044 113 374,233 0.000 30 Kweneng East Lephephe 307 987 0.256 0.047 0.218 0.036 253 374,233 0.001 31 Kweneng West Kaudwane 219 988 0.442 0.079 0.349 0.044 437 374,233 0.001 30 Kweneng East Boatlaname 491 1,678 0.326 0.059 0.218 0.036 547 374,233 0.001 31 Kweneng West Diphuduhudu 94 497 0.539 0.082 0.349 0.044 268 374,233 0.001 30 Kweneng East Metsimotlhaba 2,358 8,328 0.121 0.023 0.218 0.036 1,008 374,233 0.003 31 Kweneng West Maratshwane 73 380 0.475 0.083 0.349 0.044 181 374,233 0.000 30 Kweneng East Mmankgodi 1,367 6,389 0.288 0.049 0.218 0.036 1,840 374,233 0.005 40 Kgatleng Mochudi 10,602 42,545 0.222 0.037 0.205 0.035 9,445 374,233 0.025 30 Kweneng East Mmopane 4,251 14,635 0.119 0.02 0.218 0.036 1,742 374,233 0.005 40 Kgatleng Pilane Station 434 1,540 0.138 0.029 0.205 0.035 213 374,233 0.001 30 Kweneng East Mogoditshane 14,532 44,434 0.068 0.015 0.218 0.036 3,022 374,233 0.008 40 Kgatleng Rasesa 1,011 3,403 0.181 0.036 0.205 0.035 616 374,233 0.002 30 Kweneng East Shadishadi 297 1,149 0.351 0.076 0.218 0.036 403 374,233 0.001 40 Kgatleng Bokaa 1,490 5,885 0.194 0.037 0.205 0.035 1,142 374,233 0.003 30 Kweneng East Sojwe 844 3,269 0.298 0.054 0.218 0.036 974 374,233 0.003 40 Kgatleng Morwa 949 3,396 0.134 0.026 0.205 0.035 455 374,233 0.001 30 Kweneng East Thamaga 4,962 21,294 0.285 0.043 0.218 0.036 6,069 374,233 0.016 40 Kgatleng Matebeleng 637 2,082 0.077 0.028 0.205 0.035 160 374,233 0.000 30 Kweneng East Mmanoko 223 1,025 0.291 0.058 0.218 0.036 298 374,233 0.001 40 Kgatleng Oodi 1,437 4,780 0.114 0.023 0.205 0.035 545 374,233 0.001 30 Kweneng East Kubung 207 1,056 0.467 0.067 0.218 0.036 493 374,233 0.001 40 Kgatleng Modipane 659 2,639 0.284 0.054 0.205 0.035 749 374,233 0.002 30 Kweneng East Losilakgokong 50 192 0.271 0.086 0.218 0.036 52 374,233 0.000 40 Kgatleng Mabalane 249 725 0.108 0.034 0.205 0.035 78 374,233 0.000 30 Kweneng East Kgope 76 350 0.462 0.087 0.218 0.036 162 374,233 0.000 40 Kgatleng Sikwane 342 1,193 0.205 0.042 0.205 0.035 245 374,233 0.001 30 Kweneng East Mogonono 26 144 0.344 0.119 0.218 0.036 50 374,233 0.000 40 Kgatleng mmathubudukwane 657 2,136 0.191 0.04 0.205 0.035 408 374,233 0.001

32 Mapping Poverty in Botswana 2010 Statistics Botswana Statistics Botswana Mapping Poverty in Botswana 2010 33 Table 12.2: Estimated Disaggregated Poverty Rates at Village Level cont’d Table 12.2: Estimated Disaggregated Poverty Rates at Village Level cont’d Census District Code Census District Village Name Number of Households Population Poverty Rate Standard Error of Poverty Rate Estimated District Poverty Standard Error of Estimated District Poverty Estimated Poor in Village Estimated Poor - National Proportion of Poor in Village Census District Code Census District Village Name Number of Households Population Poverty Rate Standard Error of Poverty Rate Estimated District Poverty Standard Error of Estimated District Poverty Estimated Poor in Village Estimated Poor - National Proportion of Poor in Village 40 Kgatleng Ramonaka 138 545 0.222 0.059 0.205 0.035 121 374,233 0.000 50 Serowe/Palapye Mokgware 75 266 0.371 0.071 0.269 0.03 99 374,233 0.000 40 Kgatleng Malolwane 752 2,279 0.209 0.037 0.205 0.035 476 374,233 0.001 50 Serowe/Palapye Motshegaletau 208 906 0.435 0.061 0.269 0.03 394 374,233 0.001 40 Kgatleng Oliphants Drift 332 780 0.169 0.048 0.205 0.035 132 374,233 0.000 50 Serowe/Palapye Malatswai 162 919 0.534 0.073 0.269 0.03 491 374,233 0.001 40 Kgatleng Artesia 909 2,742 0.211 0.046 0.205 0.035 579 374,233 0.002 50 Serowe/Palapye Mokhungwana 60 248 0.311 0.073 0.269 0.03 77 374,233 0.000 40 Kgatleng Malotwana Siding 256 801 0.183 0.046 0.205 0.035 147 374,233 0.000 50 Serowe/Palapye Serule 693 2,276 0.246 0.04 0.269 0.03 560 374,233 0.001 40 Kgatleng Leshibitse 270 732 0.134 0.042 0.205 0.035 98 374,233 0.000 50 Serowe/Palapye Moreomabele 107 425 0.377 0.073 0.269 0.03 160 374,233 0.000 40 Kgatleng Ramotlabaki 94 237 0.234 0.081 0.205 0.035 55 374,233 0.000 50 Serowe/Palapye Gojwane 250 1,169 0.478 0.078 0.269 0.03 559 374,233 0.001 40 Kgatleng Kgomodiatshaba 156 508 0.292 0.081 0.205 0.035 148 374,233 0.000 50 Serowe/Palapye Manaledi 7 41 0.344 0.181 0.269 0.03 14 374,233 0.000 40 Kgatleng Dikgonnye 147 516 0.353 0.07 0.205 0.035 182 374,233 0.000 50 Serowe/Palapye Dimajwe 244 1,161 0.529 0.064 0.269 0.03 614 374,233 0.002 40 Kgatleng Khurutshe 41 84 0.186 0.08 0.205 0.035 16 374,233 0.000 50 Serowe/Palapye Majwanaadipitse 118 649 0.474 0.072 0.269 0.03 308 374,233 0.001 40 Kgatleng Dikwididi 119 407 0.21 0.068 0.205 0.035 85 374,233 0.000 50 Serowe/Palapye Sehunou 153 840 0.604 0.066 0.269 0.03 507 374,233 0.001 50 Serowe/Palapye Serowe 14,077 51,739 0.233 0.031 0.269 0.03 12,055 374,233 0.032 51 Central Mahalapye Mahalapye 11,290 42,990 0.156 0.036 0.208 0.038 6,706 374,233 0.018 50 Serowe/Palapye Palapye 10,045 35,213 0.159 0.026 0.269 0.03 5,599 374,233 0.015 51 Central Mahalapye Palla Road 399 1,501 0.183 0.051 0.208 0.038 275 374,233 0.001 50 Serowe/Palapye Lecheng 759 2,944 0.361 0.061 0.269 0.03 1,063 374,233 0.003 51 Central Mahalapye Mookane 676 2,894 0.265 0.054 0.208 0.038 767 374,233 0.002 50 Serowe/Palapye Moremi 150 615 0.352 0.075 0.269 0.03 216 374,233 0.001 51 Central Mahalapye Mmaphashalala 263 1,185 0.258 0.063 0.208 0.038 306 374,233 0.001 50 Serowe/Palapye Malaka 49 207 0.352 0.085 0.269 0.03 73 374,233 0.000 51 Central Mahalapye Taupye 144 588 0.277 0.07 0.208 0.038 163 374,233 0.000 50 Serowe/Palapye Mogapi 681 2,060 0.259 0.047 0.269 0.03 534 374,233 0.001 51 Central Mahalapye Dibete 389 1,343 0.292 0.067 0.208 0.038 392 374,233 0.001 50 Serowe/Palapye Mogapinyana 471 1,708 0.319 0.055 0.269 0.03 545 374,233 0.001 51 Central Mahalapye Dovedale 226 892 0.258 0.06 0.208 0.038 230 374,233 0.001 50 Serowe/Palapye Kgagodi 480 1,609 0.23 0.04 0.269 0.03 370 374,233 0.001 51 Central Mahalapye Kudumatse 430 1,649 0.298 0.058 0.208 0.038 491 374,233 0.001 50 Serowe/Palapye Maunatlala 1,211 4,364 0.219 0.035 0.269 0.03 956 374,233 0.003 51 Central Mahalapye Makwate 408 1,481 0.316 0.063 0.208 0.038 468 374,233 0.001 50 Serowe/Palapye Tamasane 358 1,112 0.252 0.053 0.269 0.03 280 374,233 0.001 51 Central Mahalapye Shakwe 314 1,146 0.194 0.051 0.208 0.038 222 374,233 0.001 50 Serowe/Palapye Diloro 141 536 0.344 0.062 0.269 0.03 184 374,233 0.000 51 Central Mahalapye Moshopha 451 1,533 0.175 0.053 0.208 0.038 268 374,233 0.001 50 Serowe/Palapye Lesenepole 623 2,725 0.447 0.076 0.269 0.03 1,218 374,233 0.003 51 Central Mahalapye Machaneng 659 2,497 0.278 0.054 0.208 0.038 694 374,233 0.002 50 Serowe/Palapye Mosweu 146 454 0.205 0.056 0.269 0.03 93 374,233 0.000 51 Central Mahalapye Sefhare 1,228 5,036 0.255 0.052 0.208 0.038 1,284 374,233 0.003 50 Serowe/Palapye Mokokwana 99 347 0.381 0.074 0.269 0.03 132 374,233 0.000 51 Central Mahalapye Chadibe 703 3,149 0.291 0.064 0.208 0.038 916 374,233 0.002 50 Serowe/Palapye Seolwane 392 1,350 0.284 0.057 0.269 0.03 383 374,233 0.001 51 Central Mahalapye Bonwapitse 169 627 0.231 0.062 0.208 0.038 145 374,233 0.000 50 Serowe/Palapye Lerala 1,682 6,948 0.363 0.062 0.269 0.03 2,522 374,233 0.007 51 Central Mahalapye Mmutlana 240 910 0.227 0.05 0.208 0.038 207 374,233 0.001 50 Serowe/Palapye Majwaneng 456 1,969 0.398 0.059 0.269 0.03 784 374,233 0.002 51 Central Mahalapye Mokobeng 488 1,969 0.373 0.075 0.208 0.038 734 374,233 0.002 50 Serowe/Palapye Ratholo 599 2,106 0.292 0.053 0.269 0.03 615 374,233 0.002 51 Central Mahalapye Ngwapa 94 448 0.317 0.077 0.208 0.038 142 374,233 0.000 50 Serowe/Palapye Moeng 154 353 0.048 0.027 0.269 0.03 17 374,233 0.000 51 Central Mahalapye Tumasera/Seleka 164 330 0.034 0.028 0.208 0.038 11 374,233 0.000 50 Serowe/Palapye Gootau 357 1,368 0.399 0.068 0.269 0.03 546 374,233 0.001 51 Central Mahalapye Ramokgonami 896 4,064 0.31 0.062 0.208 0.038 1,260 374,233 0.003 50 Serowe/Palapye Goo-Sekgweng 123 561 0.346 0.072 0.269 0.03 194 374,233 0.001 51 Central Mahalapye Maape 330 1,330 0.257 0.056 0.208 0.038 342 374,233 0.001 50 Serowe/Palapye Matlhakola 192 935 0.371 0.058 0.269 0.03 347 374,233 0.001 51 Central Mahalapye Pilikwe 398 1,492 0.165 0.045 0.208 0.038 246 374,233 0.001 50 Serowe/Palapye Topisi 527 1,910 0.319 0.051 0.269 0.03 609 374,233 0.002 51 Central Mahalapye Mhalapitsa 236 1,012 0.241 0.064 0.208 0.038 244 374,233 0.001 50 Serowe/Palapye Paje 668 2,523 0.294 0.044 0.269 0.03 742 374,233 0.002 51 Central Mahalapye Borotsi 189 926 0.319 0.094 0.208 0.038 295 374,233 0.001 50 Serowe/Palapye Mabeleapudi 593 2,421 0.283 0.047 0.269 0.03 685 374,233 0.002 51 Central Mahalapye Tewane 177 600 0.153 0.059 0.208 0.038 92 374,233 0.000 50 Serowe/Palapye Tshimoyapula 391 1,609 0.406 0.065 0.269 0.03 653 374,233 0.002 51 Central Mahalapye Kalamare 715 2,490 0.216 0.051 0.208 0.038 538 374,233 0.001 50 Serowe/Palapye Mmashoro 538 2,699 0.446 0.051 0.269 0.03 1,204 374,233 0.003 51 Central Mahalapye Mokgenene 131 702 0.399 0.094 0.208 0.038 280 374,233 0.001 50 Serowe/Palapye Mogorosi 548 2,799 0.455 0.051 0.269 0.03 1,274 374,233 0.003 51 Central Mahalapye Shoshong 2,941 10,419 0.162 0.037 0.208 0.038 1,688 374,233 0.005 50 Serowe/Palapye Thabala 766 3,216 0.404 0.058 0.269 0.03 1,299 374,233 0.003 51 Central Mahalapye Moralane 183 915 0.319 0.069 0.208 0.038 292 374,233 0.001 50 Serowe/Palapye Moiyabana 1,165 4,394 0.335 0.045 0.269 0.03 1,472 374,233 0.004 51 Central Mahalapye Mosolotshane 573 2,413 0.277 0.05 0.208 0.038 668 374,233 0.002 50 Serowe/Palapye Gamabuo 139 738 0.433 0.067 0.269 0.03 320 374,233 0.001 51 Central Mahalapye Kodibeleng 449 1,720 0.234 0.054 0.208 0.038 402 374,233 0.001 50 Serowe/Palapye Radisele 621 2,690 0.358 0.05 0.269 0.03 963 374,233 0.003 51 Central Mahalapye Ikongwe 124 516 0.35 0.078 0.208 0.038 181 374,233 0.000 50 Serowe/Palapye Mogome 127 559 0.359 0.064 0.269 0.03 201 374,233 0.001 51 Central Mahalapye Poloka 172 815 0.226 0.061 0.208 0.038 184 374,233 0.000

34 Mapping Poverty in Botswana 2010 Statistics Botswana Statistics Botswana Mapping Poverty in Botswana 2010 35 Table 12.2: Estimated Disaggregated Poverty Rates at Village Level cont’d Table 12.2: Estimated Disaggregated Poverty Rates at Village Level cont’d Census District Code Census District Village Name Number of Households Population Poverty Rate Standard Error of Poverty Rate Estimated District Poverty Standard Error of Estimated District Poverty Estimated Poor in Village Estimated Poor - National Proportion of Poor in Village Census District Code Census District Village Name Number of Households Population Poverty Rate Standard Error of Poverty Rate Estimated District Poverty Standard Error of Estimated District Poverty Estimated Poor in Village Estimated Poor - National Proportion of Poor in Village

51 Central Mahalapye Otse 412 1,763 0.288 0.066 0.208 0.038 508 374,233 0.001 54 Central Tutume Tutume 4,378 16,147 0.21 0.036 0.221 0.034 3,391 374,233 0.009 51 Central Mahalapye Matlhaka 154 736 0.234 0.064 0.208 0.038 172 374,233 0.000 54 Central Tutume Matobo 429 1,419 0.192 0.04 0.221 0.034 272 374,233 0.001 51 Central Mahalapye Mokoswane 136 540 0.293 0.079 0.208 0.038 158 374,233 0.000 54 Central Tutume Goshwe 368 1,538 0.243 0.056 0.221 0.034 374 374,233 0.001 51 Central Mahalapye Seleka 897 3,991 0.271 0.06 0.208 0.038 1,082 374,233 0.003 54 Central Tutume Mosetse 668 2,061 0.129 0.035 0.221 0.034 266 374,233 0.001 52 Central Bobonong Bobonong 5,501 20,217 0.312 0.046 0.346 0.043 6,308 374,233 0.017 54 Central Tutume Nswazwi 593 2,128 0.238 0.05 0.221 0.034 506 374,233 0.001 52 Central Bobonong Tsetsebjwe 1,079 4,347 0.495 0.06 0.346 0.043 2,152 374,233 0.006 54 Central Tutume Nshakashokwe 684 2,318 0.174 0.037 0.221 0.034 403 374,233 0.001 52 Central Bobonong Tshokwe 252 1,107 0.472 0.07 0.346 0.043 523 374,233 0.001 54 Central Tutume Sebina 958 3,504 0.209 0.046 0.221 0.034 732 374,233 0.002 52 Central Bobonong Mathathane 701 2,653 0.416 0.06 0.346 0.043 1,104 374,233 0.003 54 Central Tutume Marobela 488 1,782 0.221 0.046 0.221 0.034 394 374,233 0.001 52 Central Bobonong Lentselemoriti 64 192 0.342 0.091 0.346 0.043 66 374,233 0.000 54 Central Tutume Mathangwane 1,545 5,534 0.199 0.036 0.221 0.034 1,101 374,233 0.003 52 Central Bobonong Semolale 435 1,358 0.191 0.055 0.346 0.043 259 374,233 0.001 54 Central Tutume Marapong 645 2,129 0.205 0.046 0.221 0.034 436 374,233 0.001 52 Central Bobonong Kobojango 494 1,898 0.401 0.064 0.346 0.043 761 374,233 0.002 54 Central Tutume Chadibe 1,114 4,538 0.235 0.045 0.221 0.034 1,066 374,233 0.003 52 Central Bobonong Lepokole 239 956 0.495 0.065 0.346 0.043 473 374,233 0.001 54 Central Tutume Matsitama 538 1,969 0.203 0.052 0.221 0.034 400 374,233 0.001 52 Central Bobonong Molalatau 792 2,700 0.324 0.058 0.346 0.043 875 374,233 0.002 54 Central Tutume Borolong 1,310 4,908 0.195 0.039 0.221 0.034 957 374,233 0.003 52 Central Bobonong Tobane 727 2,653 0.354 0.054 0.346 0.043 939 374,233 0.003 54 Central Tutume Dukwi 1,500 5,296 0.247 0.047 0.221 0.034 1,308 374,233 0.003 52 Central Bobonong Sefophe 1,773 6,833 0.371 0.049 0.346 0.043 2,535 374,233 0.007 54 Central Tutume Mafongo/Hubona 294 1,140 0.267 0.054 0.221 0.034 304 374,233 0.001 52 Central Bobonong Mabolowe 202 704 0.292 0.054 0.346 0.043 206 374,233 0.001 54 Central Tutume Mokubilo 424 1,754 0.333 0.063 0.221 0.034 584 374,233 0.002 52 Central Bobonong Mmadinare 3,895 13,353 0.266 0.042 0.346 0.043 3,552 374,233 0.009 54 Central Tutume Tonota 5,316 20,372 0.194 0.033 0.221 0.034 3,952 374,233 0.011 52 Central Bobonong Robelela 217 810 0.315 0.072 0.346 0.043 255 374,233 0.001 54 Central Tutume Borotsi 505 1,773 0.154 0.041 0.221 0.034 273 374,233 0.001 52 Central Bobonong Damochojena 223 975 0.527 0.075 0.346 0.043 514 374,233 0.001 54 Central Tutume Mandunyane 1,006 4,051 0.261 0.048 0.221 0.034 1,057 374,233 0.003 52 Central Bobonong Motlhabaneng 336 1,422 0.456 0.064 0.346 0.043 648 374,233 0.002 54 Central Tutume Natale 302 1,227 0.257 0.047 0.221 0.034 315 374,233 0.001 52 Central Bobonong Moletemane 361 1,537 0.495 0.071 0.346 0.043 761 374,233 0.002 54 Central Tutume Shashe/Semotswane 826 3,172 0.267 0.048 0.221 0.034 847 374,233 0.002 53 Central Boteti Letlhakane 6,069 23,013 0.281 0.042 0.331 0.071 6,467 374,233 0.017 54 Central Tutume Shashe-Mooke 1,010 3,825 0.261 0.046 0.221 0.034 998 374,233 0.003 53 Central Boteti Mosu 597 2,029 0.272 0.057 0.331 0.071 552 374,233 0.001 54 Central Tutume Gweta 1,777 6,344 0.197 0.036 0.221 0.034 1,250 374,233 0.003 53 Central Boteti Mmatshumo 443 1,566 0.384 0.056 0.331 0.071 601 374,233 0.002 54 Central Tutume Zoroga 305 1,359 0.386 0.074 0.221 0.034 525 374,233 0.001 53 Central Boteti Mopipi 1,161 4,598 0.358 0.054 0.331 0.071 1,646 374,233 0.004 54 Central Tutume Lepashe 107 564 0.32 0.079 0.221 0.034 180 374,233 0.000 53 Central Boteti Xhumo 557 2,256 0.313 0.056 0.331 0.071 706 374,233 0.002 54 Central Tutume Mmanxotae 129 702 0.421 0.073 0.221 0.034 296 374,233 0.001 53 Central Boteti Kedia 165 941 0.544 0.069 0.331 0.071 512 374,233 0.001 54 Central Tutume Sepako 174 709 0.4 0.072 0.221 0.034 284 374,233 0.001 53 Central Boteti Tsienyane/Rakops 1,598 6,577 0.362 0.052 0.331 0.071 2,381 374,233 0.006 54 Central Tutume Semitwe 164 742 0.295 0.059 0.221 0.034 219 374,233 0.001 53 Central Boteti Toromoja 334 1,142 0.253 0.055 0.331 0.071 289 374,233 0.001 54 Central Tutume Makobo 46 220 0.168 0.079 0.221 0.034 37 374,233 0.000 53 Central Boteti Moreomaoto 178 648 0.313 0.065 0.331 0.071 203 374,233 0.001 54 Central Tutume Mabesekwa 550 2,155 0.324 0.061 0.221 0.034 698 374,233 0.002 53 Central Boteti Makalamabedi 499 1,698 0.313 0.063 0.331 0.071 531 374,233 0.001 54 Central Tutume Jamataka 151 631 0.347 0.067 0.221 0.034 219 374,233 0.001 53 Central Boteti Khwee 172 952 0.669 0.071 0.331 0.071 637 374,233 0.002 54 Central Tutume Tshokatshaa 73 383 0.377 0.086 0.221 0.034 144 374,233 0.000 53 Central Boteti Kumaga 257 874 0.339 0.068 0.331 0.071 296 374,233 0.001 54 Central Tutume Makuta 273 988 0.143 0.04 0.221 0.034 141 374,233 0.000 53 Central Boteti Motopi 276 1,217 0.41 0.067 0.331 0.071 499 374,233 0.001 54 Central Tutume Maposa 89 312 0.326 0.081 0.221 0.034 102 374,233 0.000 53 Central Boteti Mmadikola 258 1,047 0.336 0.062 0.331 0.071 352 374,233 0.001 54 Central Tutume Kutamogree 487 2,018 0.28 0.055 0.221 0.034 565 374,233 0.002 53 Central Boteti Mokoboxane 274 1,240 0.362 0.064 0.331 0.071 449 374,233 0.001 54 Central Tutume Mmeya 155 766 0.442 0.08 0.221 0.034 339 374,233 0.001 53 Central Boteti Xere 106 396 0.53 0.086 0.331 0.071 210 374,233 0.001 60 North East Mbalambi 245 1,017 0.274 0.06 0.188 0.036 279 374,233 0.001 54 Central Tutume Maitengwe 1,550 6,092 0.248 0.045 0.221 0.034 1,511 374,233 0.004 60 North East Gambule 117 458 0.194 0.067 0.188 0.036 89 374,233 0.000 54 Central Tutume Nata 1,574 5,588 0.288 0.036 0.221 0.034 1,609 374,233 0.004 60 North East Gulubane 238 813 0.115 0.048 0.188 0.036 93 374,233 0.000 54 Central Tutume Dagwi 261 1,020 0.231 0.056 0.221 0.034 236 374,233 0.001 60 North East Gungwe 104 498 0.203 0.067 0.188 0.036 101 374,233 0.000 54 Central Tutume Nkange 946 3,494 0.227 0.045 0.221 0.034 793 374,233 0.002 60 North East Jackalas 1 331 1,114 0.184 0.046 0.188 0.036 205 374,233 0.001 54 Central Tutume Senete 682 2,681 0.252 0.046 0.221 0.034 676 374,233 0.002 60 North East Jackalas 2 269 1,046 0.298 0.059 0.188 0.036 312 374,233 0.001 54 Central Tutume Changate 272 1,090 0.198 0.063 0.221 0.034 216 374,233 0.001 60 North East Kalakamati 244 823 0.168 0.053 0.188 0.036 138 374,233 0.000

36 Mapping Poverty in Botswana 2010 Statistics Botswana Statistics Botswana Mapping Poverty in Botswana 2010 37 Table 12.2: Estimated Disaggregated Poverty Rates at Village Level cont’d Table 12.2: Estimated Disaggregated Poverty Rates at Village Level cont’d Census District Code Census District Village Name Number of Households Population Poverty Rate Standard Error of Poverty Rate Estimated District Poverty Standard Error of Estimated District Poverty Estimated Poor in Village Estimated Poor - National Proportion of Poor in Village Census District Code Census District Village Name Number of Households Population Poverty Rate Standard Error of Poverty Rate Estimated District Poverty Standard Error of Estimated District Poverty Estimated Poor in Village Estimated Poor - National Proportion of Poor in Village

60 North East Letsholathebe 112 492 0.329 0.073 0.188 0.036 162 374,233 0.000 70 Ngamiland East Matlapana 210 1,018 0.424 0.05 0.306 0.034 432 374,233 0.001 60 North East Makaleng 297 1,117 0.18 0.048 0.188 0.036 201 374,233 0.001 70 Ngamiland East Phuduhudu 125 571 0.397 0.075 0.306 0.034 227 374,233 0.001 60 North East Mulambakwena 286 964 0.146 0.044 0.188 0.036 141 374,233 0.000 70 Ngamiland East Komana 62 215 0.266 0.084 0.306 0.034 57 374,233 0.000 60 North East Mambo 45 156 0.113 0.074 0.188 0.036 18 374,233 0.000 70 Ngamiland East Sankuyo 72 382 0.446 0.087 0.306 0.034 170 374,233 0.000 60 North East Mapoka 372 1,393 0.231 0.045 0.188 0.036 322 374,233 0.001 70 Ngamiland East Mababe 65 217 0.382 0.092 0.306 0.034 83 374,233 0.000 60 North East Masingwaneng 128 469 0.244 0.057 0.188 0.036 114 374,233 0.000 70 Ngamiland East Chanoga 81 341 0.331 0.072 0.306 0.034 113 374,233 0.000 60 North East Masunga 1,515 4,287 0.069 0.019 0.188 0.036 296 374,233 0.001 70 Ngamiland East Bodibeng 78 383 0.433 0.076 0.306 0.034 166 374,233 0.000 60 North East Matsiloje 764 2,632 0.173 0.034 0.188 0.036 455 374,233 0.001 70 Ngamiland East Botlhatlogo 156 689 0.433 0.073 0.306 0.034 298 374,233 0.001 60 North East Moroka 403 1,644 0.246 0.051 0.188 0.036 404 374,233 0.001 70 Ngamiland East Semboyo 68 389 0.461 0.083 0.306 0.034 179 374,233 0.000 60 North East Mosojane 300 1,173 0.256 0.056 0.188 0.036 300 374,233 0.001 70 Ngamiland East Kgakge/Makakung 25 137 0.445 0.123 0.306 0.034 61 374,233 0.000 60 North East Nlakhwane 399 1,496 0.216 0.051 0.188 0.036 323 374,233 0.001 70 Ngamiland East Habu 101 488 0.27 0.067 0.306 0.034 132 374,233 0.000 60 North East Ramokgwebana 389 1,490 0.202 0.046 0.188 0.036 301 374,233 0.001 71 Ngamiland West Seronga 731 2,879 0.447 0.053 0.375 0.049 1,287 374,233 0.003 60 North East Sechele 123 501 0.265 0.065 0.188 0.036 133 374,233 0.000 71 Ngamiland West Ngarange 340 1,357 0.325 0.054 0.375 0.049 441 374,233 0.001 60 North East Sekakangwe 187 657 0.321 0.074 0.188 0.036 211 374,233 0.001 71 Ngamiland West Beetsha 343 1,530 0.478 0.066 0.375 0.049 731 374,233 0.002 60 North East Senyawe 340 1,269 0.232 0.043 0.188 0.036 294 374,233 0.001 71 Ngamiland West Gonutsuga 196 908 0.427 0.062 0.375 0.049 388 374,233 0.001 60 North East Siviya 244 1,016 0.292 0.055 0.188 0.036 297 374,233 0.001 71 Ngamiland West Mokgacha 93 492 0.519 0.091 0.375 0.049 255 374,233 0.001 60 North East Tati Siding 2,040 7,501 0.116 0.025 0.188 0.036 870 374,233 0.002 71 Ngamiland West Qangwa 223 1,017 0.435 0.064 0.375 0.049 442 374,233 0.001 60 North East Themashanga 370 1,664 0.382 0.06 0.188 0.036 636 374,233 0.002 71 Ngamiland West Xakao 324 1,442 0.387 0.061 0.375 0.049 558 374,233 0.001 60 North East Tsamaya 549 1,957 0.196 0.045 0.188 0.036 384 374,233 0.001 71 Ngamiland West Kauxwhi 336 1,576 0.393 0.056 0.375 0.049 619 374,233 0.002 60 North East Tshesebe 575 1,950 0.167 0.039 0.188 0.036 326 374,233 0.001 71 Ngamiland West Etsha 6 1,026 4,902 0.41 0.048 0.375 0.049 2,010 374,233 0.005 60 North East Zwenshambe 399 1,491 0.184 0.048 0.188 0.036 274 374,233 0.001 71 Ngamiland West Ikoga 311 1,207 0.322 0.059 0.375 0.049 389 374,233 0.001 60 North East Butale 144 528 0.183 0.06 0.188 0.036 97 374,233 0.000 71 Ngamiland West Sepopa 544 2,206 0.39 0.046 0.375 0.049 860 374,233 0.002 60 North East Masukwane 161 625 0.149 0.056 0.188 0.036 93 374,233 0.000 71 Ngamiland West Nxamasere 325 1,494 0.383 0.055 0.375 0.049 572 374,233 0.002 60 North East Shashe Bridge 200 829 0.059 0.032 0.188 0.036 49 374,233 0.000 71 Ngamiland West Shakawe 1,592 6,780 0.377 0.05 0.375 0.049 2,556 374,233 0.007 60 North East Matshelagabedi 702 3,069 0.206 0.043 0.188 0.036 632 374,233 0.002 71 Ngamiland West Nxaunxau 100 459 0.443 0.061 0.375 0.049 203 374,233 0.001 60 North East Ditladi 398 1,485 0.184 0.047 0.188 0.036 273 374,233 0.001 71 Ngamiland West Mohembo West 435 1,920 0.496 0.054 0.375 0.049 952 374,233 0.003 60 North East Botalaote 55 186 0.127 0.068 0.188 0.036 24 374,233 0.000 71 Ngamiland West Etsha 13 603 2,646 0.359 0.057 0.375 0.049 950 374,233 0.003 60 North East Toteng 72 265 0.179 0.075 0.188 0.036 47 374,233 0.000 71 Ngamiland West Nokaneng 712 3,243 0.246 0.054 0.375 0.049 798 374,233 0.002 60 North East Matenge 113 406 0.165 0.061 0.188 0.036 67 374,233 0.000 71 Ngamiland West Gani 148 677 0.484 0.076 0.375 0.049 328 374,233 0.001 60 North East Vukwi 87 363 0.328 0.069 0.188 0.036 119 374,233 0.000 71 Ngamiland West Tubu 130 583 0.238 0.068 0.375 0.049 139 374,233 0.000 60 North East Pole 71 277 0.215 0.08 0.188 0.036 60 374,233 0.000 71 Ngamiland West Gumare 2,001 8,212 0.328 0.045 0.375 0.049 2,694 374,233 0.007 60 North East Kgari 139 524 0.2 0.061 0.188 0.036 105 374,233 0.000 71 Ngamiland West Xaxa 102 508 0.526 0.095 0.375 0.049 267 374,233 0.001 60 North East Mowana 103 476 0.207 0.062 0.188 0.036 99 374,233 0.000 71 Ngamiland West Mohembo East 113 543 0.377 0.079 0.375 0.049 205 374,233 0.001 60 North East Mabudzane 113 556 0.261 0.062 0.188 0.036 145 374,233 0.000 71 Ngamiland West Tobere 136 473 0.243 0.062 0.375 0.049 115 374,233 0.000 60 North East Matopi 84 347 0.339 0.085 0.188 0.036 118 374,233 0.000 71 Ngamiland West Sekondomboro 172 610 0.249 0.058 0.375 0.049 152 374,233 0.000 60 North East Patayamatebele 92 342 0.207 0.069 0.188 0.036 71 374,233 0.000 71 Ngamiland West Samochema 244 1,133 0.491 0.058 0.375 0.049 556 374,233 0.001 70 Ngamiland East Maun 14,349 58,877 0.291 0.034 0.306 0.034 17,133 374,233 0.046 71 Ngamiland West Xhauga 101 780 0.274 0.087 0.375 0.049 214 374,233 0.001 70 Ngamiland East Makalamabedi 582 1,831 0.27 0.044 0.306 0.034 494 374,233 0.001 71 Ngamiland West Mogomotho 189 793 0.275 0.072 0.375 0.049 218 374,233 0.001 70 Ngamiland East Shorobe 651 2,473 0.408 0.045 0.306 0.034 1,009 374,233 0.003 71 Ngamiland West Eretsha 245 909 0.426 0.074 0.375 0.049 387 374,233 0.001 70 Ngamiland East Toteng 724 2,452 0.268 0.045 0.306 0.034 657 374,233 0.002 71 Ngamiland West Chukumuchu 37 147 0.423 0.106 0.375 0.049 62 374,233 0.000 70 Ngamiland East Tsao 607 2,644 0.316 0.05 0.306 0.034 836 374,233 0.002 71 Ngamiland West Gudingwa 136 710 0.516 0.094 0.375 0.049 366 374,233 0.001 70 Ngamiland East Sehithwa 1,208 4,428 0.289 0.039 0.306 0.034 1,280 374,233 0.003 71 Ngamiland West Etsha 13 280 1,272 0.3 0.065 0.375 0.049 382 374,233 0.001 70 Ngamiland East Kareng 348 1,358 0.358 0.05 0.306 0.034 486 374,233 0.001 71 Ngamiland West Kajaja 31 137 0.512 0.115 0.375 0.049 70 374,233 0.000 70 Ngamiland East Matsaudi/Sakapnae 93 334 0.366 0.069 0.306 0.034 122 374,233 0.000 72 Chobe Kachikau 359 1,197 0.411 0.06 0.195 0.029 492 374,233 0.001

38 Mapping Poverty in Botswana 2010 Statistics Botswana Statistics Botswana Mapping Poverty in Botswana 2010 39 Table 12.2: Estimated Disaggregated Poverty Rates at Village Level cont’d Table 12.2: Estimated Disaggregated Poverty Rates at Village Level cont’d Census District Code Census District Village Name Number of Households Population Poverty Rate Standard Error of Poverty Rate Estimated District Poverty Standard Error of Estimated District Poverty Estimated Poor in Village Estimated Poor - National Proportion of Poor in Village Census District Code Census District Village Name Number of Households Population Poverty Rate Standard Error of Poverty Rate Estimated District Poverty Standard Error of Estimated District Poverty Estimated Poor in Village Estimated Poor - National Proportion of Poor in Village

72 Chobe Kasane 2,872 8,217 0.13 0.022 0.195 0.029 1,068 374,233 0.003 90 Kgalagadi South Bokspits 185 622 0.173 0.048 0.233 0.036 108 374,233 0.000 72 Chobe Kazungula 1,172 3,760 0.196 0.031 0.195 0.029 737 374,233 0.002 90 Kgalagadi South Struizendam 96 467 0.336 0.072 0.233 0.036 157 374,233 0.000 72 Chobe Kavimba 175 533 0.191 0.052 0.195 0.029 102 374,233 0.000 90 Kgalagadi South Bray 208 990 0.336 0.065 0.233 0.036 333 374,233 0.001 72 Chobe Lesoma 184 652 0.217 0.054 0.195 0.029 141 374,233 0.000 90 Kgalagadi South Phepheng/Draaihoek 219 962 0.316 0.066 0.233 0.036 304 374,233 0.001 72 Chobe Muchinje/Mabele 212 747 0.251 0.052 0.195 0.029 187 374,233 0.001 90 Kgalagadi South Mabuelo 126 501 0.317 0.067 0.233 0.036 159 374,233 0.000 72 Chobe Parakarungu 246 987 0.358 0.069 0.195 0.029 353 374,233 0.001 90 Kgalagadi South Khawa 153 809 0.436 0.085 0.233 0.036 353 374,233 0.001 72 Chobe Pandamatenga 609 1,943 0.189 0.039 0.195 0.029 367 374,233 0.001 90 Kgalagadi South Kokotsha 295 1,460 0.339 0.06 0.233 0.036 495 374,233 0.001 72 Chobe Satau 120 375 0.3 0.063 0.195 0.029 113 374,233 0.000 90 Kgalagadi South Maralaleng 140 593 0.255 0.064 0.233 0.036 151 374,233 0.000 73 Delta Xaxaba 49 162 0.526 0.092 0.352 0.071 85 374,233 0.000 90 Kgalagadi South Maleshe 128 502 0.206 0.061 0.233 0.036 103 374,233 0.000 73 Delta Jao 22 128 0.436 0.158 0.352 0.071 56 374,233 0.000 91 Kgalagadi North Kang 1,511 4,603 0.207 0.033 0.239 0.035 953 374,233 0.003 73 Delta Katamaga 14 36 0.05 0.068 0.352 0.071 2 374,233 0.000 91 Kgalagadi North Tshane 289 910 0.179 0.044 0.239 0.035 163 374,233 0.000 73 Delta Daonara 12 28 0.094 0.115 0.352 0.071 3 374,233 0.000 91 Kgalagadi North Hukuntsi 1,379 4,318 0.153 0.032 0.239 0.035 661 374,233 0.002 73 Delta Ditshiping 105 236 0.249 0.08 0.352 0.071 59 374,233 0.000 91 Kgalagadi North Lehututu 446 1,784 0.267 0.053 0.239 0.035 476 374,233 0.001 73 Delta Morutsha 14 54 0.369 0.206 0.352 0.071 20 374,233 0.000 91 Kgalagadi North Lokgwabe 378 1,469 0.21 0.053 0.239 0.035 308 374,233 0.001 80 Ghanzi Ghanzi 3,920 12,999 0.222 0.034 0.301 0.038 2,886 374,233 0.008 91 Kgalagadi North Monong 67 255 0.448 0.107 0.239 0.035 114 374,233 0.000 80 Ghanzi Dekar 279 1,456 0.593 0.056 0.301 0.038 863 374,233 0.002 91 Kgalagadi North Ncaang 58 218 0.291 0.087 0.239 0.035 63 374,233 0.000 80 Ghanzi Tsootsha 598 1,936 0.246 0.043 0.301 0.038 476 374,233 0.001 91 Kgalagadi North Hunhukwe 231 671 0.263 0.06 0.239 0.035 176 374,233 0.000 80 Ghanzi Karakobis 365 1,076 0.275 0.051 0.301 0.038 296 374,233 0.001 91 Kgalagadi North Zutswa 138 547 0.45 0.07 0.239 0.035 246 374,233 0.001 80 Ghanzi Chobokwane 262 996 0.3 0.067 0.301 0.038 299 374,233 0.001 91 Kgalagadi North Ngwatle 60 231 0.513 0.11 0.239 0.035 119 374,233 0.000 80 Ghanzi West Hanahai 179 744 0.39 0.072 0.301 0.038 290 374,233 0.001 91 Kgalagadi North Ukwi 133 491 0.32 0.077 0.239 0.035 157 374,233 0.000 80 Ghanzi Charles Hill 1,042 3,361 0.194 0.032 0.301 0.038 652 374,233 0.002 91 Kgalagadi North Make 91 397 0.485 0.071 0.239 0.035 193 374,233 0.001 80 Ghanzi Makunda 245 989 0.28 0.053 0.301 0.038 277 374,233 0.001 91 Kgalagadi North Inalegolo 70 332 0.574 0.089 0.239 0.035 191 374,233 0.001 80 Ghanzi Kule 326 1,032 0.269 0.048 0.301 0.038 278 374,233 0.001 91 Kgalagadi North Phuduhudu 187 618 0.236 0.067 0.239 0.035 146 374,233 0.000 80 Ghanzi Ncojane 715 2,274 0.24 0.041 0.301 0.038 546 374,233 0.001 80 Ghanzi Groote Laagte 137 897 0.646 0.064 0.301 0.038 579 374,233 0.002 80 Ghanzi New Xanagas 231 994 0.414 0.057 0.301 0.038 412 374,233 0.001 80 Ghanzi East Hanahai 136 555 0.417 0.089 0.301 0.038 231 374,233 0.001 80 Ghanzi Kacgae 155 606 0.423 0.079 0.301 0.038 256 374,233 0.001 80 Ghanzi Bere 201 758 0.407 0.059 0.301 0.038 309 374,233 0.001 80 Ghanzi Qabo 132 708 0.69 0.074 0.301 0.038 489 374,233 0.001 80 Ghanzi New Xade 196 810 0.456 0.063 0.301 0.038 369 374,233 0.001 90 Kgalagadi South Werda 708 2,972 0.288 0.051 0.233 0.036 856 374,233 0.002 90 Kgalagadi South Makopong 473 1,827 0.254 0.057 0.233 0.036 464 374,233 0.001 90 Kgalagadi South Khisa 90 401 0.302 0.093 0.233 0.036 121 374,233 0.000 90 Kgalagadi South Omaweneno 242 969 0.22 0.051 0.233 0.036 213 374,233 0.001 90 Kgalagadi South Tsabong 2,684 8,631 0.154 0.031 0.233 0.036 1,329 374,233 0.004 90 Kgalagadi South Kolonkwane 201 666 0.185 0.053 0.233 0.036 123 374,233 0.000 90 Kgalagadi South Bogogobo 94 338 0.365 0.111 0.233 0.036 123 374,233 0.000 90 Kgalagadi South Middlepits 247 736 0.139 0.039 0.233 0.036 102 374,233 0.000 90 Kgalagadi South Khuis 262 915 0.206 0.051 0.233 0.036 188 374,233 0.001 90 Kgalagadi South Gachibana 238 908 0.155 0.044 0.233 0.036 141 374,233 0.000 90 Kgalagadi South Rapples Pan 90 381 0.231 0.059 0.233 0.036 88 374,233 0.000 90 Kgalagadi South Vaalhoek 69 348 0.285 0.073 0.233 0.036 99 374,233 0.000

40 Mapping Poverty in Botswana 2010 Statistics Botswana Statistics Botswana Mapping Poverty in Botswana 2010 41 Appendix of Figures Figure 2. Beta Model Residuals of Models 1 to 3 Figure 1. Beta residuals and Fitted Values

Model 1: Comparable Variables with no interactions and no village level variables

Model 1: Comparable Variables with no interactions and no village level variables

Model 1: Comparable Variables with no interactions and no village level variables

42 Mapping Poverty in Botswana 2010 Statistics Botswana Statistics Botswana Mapping Poverty in Botswana 2010 43 Figure 3. Comparison of LNEXP and per capital LNEXP models Quartile-Quartile Plot

Poverty FigureIncid 5.e Villagence R Povertyates Incidenceat the V byil lDistrictage Level Poverty Map Estimates

Kasane

Shakawe ±

Chobe

Maun Ngamiland Nata

Tutume Sowa Sehithwa Masunga

North East Orapa Francistown Figure 3. Comparison of LNEXP and per capital LNEXP models Letlhakane Central Tonota Quartile-Quartile Plot Ghanzi

Selebi Phikwe Mamuno Serowe Bobonong Ghanzi Palapye

Mahalapye

Kang K we ne ng

Kgatleng Mochudi Jwaneng Kgalagadi Gaborone Ramotswa S o u South East th Lobatse e rn

Tsabong Legend

Poverty Incidence Rates

Under 0.20 0.20 - 0.25 0.25 - 0.30 0.30 - 0.35 0.35 - 0.40

Above 0.40

Kilometers Administrative District 0 35 70 140 210 280 Major Roads

PPreparedrepared by C abyrto gCartographicraphic Unit, Statisti cUnit,s Bots Statisticswana, Gabo rBotswana,one 2014 Gaborone 2014

44 Mapping Poverty in Botswana 2010 Statistics Botswana Statistics Botswana Mapping Poverty in Botswana 2010 45 Figure 7. Confidence intervals of village level estimates ELL Standard Method

PoveFigurerty D6. eVillagensit Povertyy at t hDensitye Vi lbyla Districtge L evel Poverty Map Estimates

Kasane

Shakawe ±

Chobe

Ngamiland Maun

Nata Tutume Masunga Sehithwa Sowa North East Orapa Francistown

Central Figure 8. Significant different pairwise comparisons of villages Ghanzi Letlhakane Tonota Selebi Phikwe within the nation Bobonong Serowe Mamuno Palapye Ghanzi

Mahalapye Kang K we ne ng

Kgatleng Molepolole Mochudi

Kgalagadi Jwaneng Gaborone Thamaga S Ramotswa o u South East th e Lobatse rn

Tsabong Legend

Poverty Density Rates

Under 100

100 - 400

400 - 1000

1000 - 4000

Above 4000

Kilometers Administrative District 0 35 70 140 210 280 Major Roads

PrepPreparedared by Cart obygra pCartographichic Unit, Statistics Unit, Botsw Statisticsana, Gabo rBotswana,one 2014 Gaborone 2014

46 Mapping Poverty in Botswana 2010 Statistics Botswana Statistics Botswana Mapping Poverty in Botswana 2010 47 Figure 9. Significant different pairwise comparisons of villages Figure 10. Significantly different pairwise comparisons of villages within the nation (Bonferroni Correction) within districts (Bonferroni Correction)

Note: The significance level has been divided by 124,750 which correspond to the total number of comparisons.

Note: Seven districts are not considered in this graphic given that they contain only one village per district.

48 Mapping Poverty in Botswana 2010 Statistics Botswana Statistics Botswana Mapping Poverty in Botswana 2010 49 Figure 11. Proportion of villages with poverty rates significantly Figure 12. Confidence intervals of village level estimates different from the district poverty rate Empirical Bayes prediction (EB) (Bonferroni Correction)

Figure 13. Confidence intervals of village level estimates (Only villages appearing in survey and census)

50 Mapping Poverty in Botswana 2010 Statistics Botswana Statistics Botswana Mapping Poverty in Botswana 2010 51 Ngamiland North East

Ghanzi

Central

Kgalagadi

kgatleng

South East